Headed to AWS re:Invent? We’re teaming up with @baseten to host a small, invite-only gathering for AI leaders before the first keynotes and sessions kick off.
☕️ RSVP today 🥐
luma.com/uf3erp0a
LangChain
LangChain is a software framework that helps facilitate the integration of large language models (LLMs) into applications. As a language model integration framework, LangChain's use-cases largely overlap with those of language models in general, including document analysis and summarization, chatbots, and code analysis. More on Wikipedia
-
LangChain
@LangChain
-
LangChain
@LangChain
We made agents the core object across LangSmith, connecting fragmented workflows for developers and enabling platform teams to manage agents across the business.
Now, teams can manage the entire agent development lifecycle from one place, using any stack, with ownership and governance across all of it.
LangChain
@LangChain
Meet the new LangSmith Agent Platform: a single place to build, run, improve, and govern every agent across your org.
Head of Product @samecrowder on:
1️⃣ Why we reorganized it
2️⃣ How the agent model works
3️⃣ What it unlocks
4️⃣ Where it’s going next
www.langchain.com/blog/we-rebuilt-langsmith-around-the-agent -
LangChain
@LangChain
This Friday, join @hwchase17 for a conversation on the future of agents, what's changing in how they're built, and what it will take to make them more capable, reliable, and scale in production.
RSVP: events.langchain.com/webinar/future-of-agents -
LangChain reposted
Liam Bush @LiamBush20
In six months, agents will be “always-on employees" capable of owning work end to end.
Agent governance will be the most important aspect of harness design going forward.
Harrison Chase
@hwchase17
as models get more capable, we see them moving from doing tasks towards pursuing goals
this is changes what an agent does, what it needs, how its built, and we interact with it
this blog covers what we see happening x.com/i/article/2122379034319577088 -
LangChain reposted
Sydney Runkle
@sydneyrunkle
the most important type of context is domain knowledge and taste
this is often delivered in the form of feedback, so review UX needs to be delightful
this enables agents to continually learn and improve, becoming increasingly a) useful and b) autonomous
Harrison Chase
@hwchase17
as agents move towards working on goals, it becomes important to design those goals correctly
wrote with @mgarimella on some lessons learned at actively! x.com/i/article/2121382906820243456 -
LangChain
@LangChain
Meet the new LangSmith Agent Platform: a single place to build, run, improve, and govern every agent across your org.
Head of Product @samecrowder on:
1️⃣ Why we reorganized it
2️⃣ How the agent model works
3️⃣ What it unlocks
4️⃣ Where it’s going next
www.langchain.com/blog/we-rebuilt-langsmith-around-the-agent -
LangChain
@LangChain
Happening tomorrow in SF!
Join our hands-on workshop exploring Managed Deep Agents on LangSmith.
You’ll get practical experience building, testing, and deploying a production-ready data analysis agent.
Spaces are limited! RSVP ⤵️
fieldevents.langchain.com/buildanddeployanagentwithmanag -
LangChain
@LangChain
"In the future, agents will advance to the point where you just give them a goal, and they go out and work to optimize that goal."
New blog from @hwchase17 on goal engineering.
Harrison Chase
@hwchase17
as agents move towards working on goals, it becomes important to design those goals correctly
wrote with @mgarimella on some lessons learned at actively! x.com/i/article/2121382906820243456 -
LangChain
@LangChain
At Interrupt NYC, the @JPMorgan Chase team took the stage to share how Jarvis, their internal agent assembly line, turned months of building into minutes, and one agent into a fleet.
⏯️ Watch the full session: youtu.be/VpHYPfvfy0oThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
There’s still so much left to build.
Building high performing agents is harder than it should be, and we're determined to change that.
If you want to join us on our mission, take a look at our open roles today. www.langchain.com/careers
LangChain
@LangChain
LangChain has raised $300M at a $3B valuation to advance our mission to enable every company to own its intelligence. www.langchain.com/blog/series-c
More than 7,000 companies already rely on LangSmith for agent runtime, improvement, and governance. We're excited to make it easier for every company to own its intelligence. -
LangChain
@LangChain
Over the past 3 years, millions of developers have used LangChain frameworks to build agents.
Today, we're putting everything we've learned into one platform: the new LangSmith.
Built for and powered by agents, supporting the entire agent development lifecycle, so you can build, run, improve, and govern every agent.
Harrison Chase
@hwchase17
Today we announced that LangChain raised $300M at a $3B valuation.
We think agents will transform the economy, and the companies that own their intelligence will build lasting competitive advantages.
That's why we're building LangSmith, the framework-agnostic platform to run, improve, and govern agents.
Still so much left to build! More on what’s next: www.langchain.com/blog/series-cThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
LangChain has raised $300M at a $3B valuation to advance our mission to enable every company to own its intelligence. www.langchain.com/blog/series-c
More than 7,000 companies already rely on LangSmith for agent runtime, improvement, and governance. We're excited to make it easier for every company to own its intelligence.
LangChain
@LangChain
Series C led by @CapitalG
New investors: GIC, @DFJgrowth, @SalesforceVC, @insightpartners, QuantumLight, + @MongoDB Ventures
Existing investors: @SapphireVC, @IVP, @Sequoia, @Benchmark, @AmplifyPartners, @Frontlinevc
events.langchain.com/webinar/future-of-agents/ -
LangChain
@LangChain
⏯️ Access all the Interrupt 26 recordings: interrupt.langchain.com/recordings
🎟️ RSVP for Interrupt 27: interrupt.langchain.com
LangChain
@LangChain
At Interrupt NYC, @hwchase17 sat down with @Modal CEO & Co-Founder @bernhardsson for a fireside chat on sandboxes, inference, and more.
Watch the full session: youtu.be/1M8vagUOdJMThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
LangSmith for Startups Spotlight: @Meet_campfire 🔥
Campfire is the AI-native ERP for finance teams, built around Ember, an accounting agent that reconciles accounts, books journal entries, and drives the close alongside the team.
LangChain
@LangChain
Campfire uses LangSmith to design and debug Ember, tracing every step of each agent run so the team can see how Ember reasons, which tools it calls, and where it goes wrong.
Campfire is hiring engineers to build Ember → campfire.ai/careers -
LangChain
@LangChain
New LangChain Academy course: Building and Deploying Managed Agents 🚀
An agent can take weeks to ship to real users because of all the infrastructure that has to run alongside it.
Led by @ishaan_lal1, learn to close that runtime gap
Enroll free: academy.langchain.com/courses/managed-agentsThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
🎉 There's still time to RSVP for tonight's Friends of LangChain Party at our SF Office!
Come hang out with the LangChain team, network with the community, and enjoy a DJ, fun activations, and a few surprises!
We’d love to celebrate with you! luma.com/i4xu7a6l -
LangChain
@LangChain
At Interrupt NYC, @hwchase17 sat down with @Modal CEO & Co-Founder @bernhardsson for a fireside chat on sandboxes, inference, and more.
Watch the full session: youtu.be/1M8vagUOdJMThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
ICYMI: We launched our State Of Agent Engineering survey.
Share your top priorities and use cases for agent engineering.
Answer for a chance to win exclusive swag!
Sydney Runkle
@sydneyrunkle
we just released our annual “state of agent eng” survey!
this data helps us get a pulse on how teams are building agents, and we share the results publicly afterwards!
takes about 5 min, and everyone who answers gets a chance to win some LC swag 😎
form.typeform.com/to/U2EAO1Lz -
LangChain reposted
Sydney Runkle
@sydneyrunkle
gemini decisions API launched today!
this API will feel very familiar if you've used jev or other decision models. gemini decisions is much faster and cheaper than an LLM, and outputs are structured + reliable :)
we have day 0 support in langchain: docs.langchain.com/oss/python/integrations/decision_model…
Sundar Pichai
@sundarpichai
Today we’re launching the Gemini Decisions API, powered by our new Gemini Decision 1 model in @GoogleAIStudio.
Instead of generating text token by token, it’s built for closed-set choices scoring up to 1,000 questions across text, images, audio, video + PDFs in a single pass. A few details:
- 3 decision types: yes/no, multiple choice (up to 255 options) + rubric scoring
- $0.10 / 1M input tokens with $0 billed for output tokens (~5x cheaper than Flash-Lite)
- leads on several benchmark evals with lowest calibration error
Google teams are already building with it, below is a look at Stitch Rapid designing a full product flow in seconds (1x speed) -
LangChain reposted
Harrison Chase
@hwchase17
we're running our annual "State of Agent Engineering" survey!
we're all just trying to figure out how to build agents together - these are a few of the questions/topics that are top of mind of us
excited to see what the community thinks!
form.typeform.com/to/U2EAO1Lz?typeform-source=t.co -
LangChain reposted
Caspar
@caspar_br
We're running a MDA workshop in SF next Tuesday! Bring a laptop and leave with a deployed agent langch.in/mda-workshop
-
LangChain reposted
Sydney Runkle
@sydneyrunkle
here's my "proactive agent engineering" talk slides!
thesis: there's a shift coming from "you give an agent tasks" to "an agent gives you tasks"
this requires a step change in memory and ux
docs.google.com/presentation/d/1VbC01MaG_YjALvag9v1NX00Gj…
Sydney Runkle
@sydneyrunkle
giving a talk at @aiDotEngineer in a few hours on my current favorite subject: proactive agents!
blog/slides coming soon 🤓 -
LangChain
@LangChain
Attending re:Invent? On December 3rd, stop by the Agent Arena, sponsored by LangChain, @baseten, and @RedisInc.
🍹 Fuel up with bites & cocktails
👥 Connect with peers
🕹️ Play & win
RSVP today: redis.io/events/agent-arena-at-aws-reinvent-2026/ -
LangChain reposted
Alex Olsen
@its_ao
I'm lucky to be surrounded by great engineers at LangChain
now you can too!
Come hang with us this Thurs + help us figure out if Make no mistakes actually works!
luma.com/x2d94ihu -
LangChain
@LangChain
Devs and agent engineers: we want your input! Take our survey to help shape the 2027 State of Agent Engineering report (and for the chance to win some limited edition LangChain swag).
💻 form.typeform.com/to/U2EAO1Lz -
LangChain
@LangChain
Amazing opportunity to learn about Managed Deep Agents in-person. langch.in/mda-workshop
Caspar
@caspar_br
We're running a MDA workshop in SF next Tuesday! Bring a laptop and leave with a deployed agent langch.in/mda-workshop
-
LangChain reposted
Prefect
@PrefectIO
git ready with us. FastMCP is going to GitHub HQ.
Wed, Dec 2, 6-9 pm PT, San Francisco.
@jlowin, @zzstoatzz, @sydneyrunkle, and @pamelafox on skills over MCP, auth, async tasks, Postgres, and MCP apps. then food & drinks!
commit to it, link in comments. -
LangChain
@LangChain
Headed to AIE Code? Don't miss @sydneyrunkle's upcoming session.
Sydney Runkle
@sydneyrunkle
giving a talk at @aiDotEngineer in a few hours on my current favorite subject: proactive agents!
blog/slides coming soon 🤓 -
LangChain
@LangChain
A @Workday agent prototype used to take 4 to 5 months, with no guarantee of reaching production.
With LangGraph, Agent Server, and LangSmith, teams can now get agents running in staging in under a day, with auth and security built in.
This is their story.This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Happening today!
Join us for an evening of talks and community discussion about @ApacheDataFusio and its growing role in modern data infrastructure.
Get more info + RSVP: luma.com/k04h0om1 -
LangChain
@LangChain
Happening this week!
Come explore a career at LangChain!
luma.com/langchain-np4d -
LangChain
@LangChain
Come for the crash courses.
Stay for the memes.
Jess Weng
@jess_weng18
More @LangChain Deep Agents workshop memes
(Retrieval & Context Engineering edition)
Retrieval: goldcast.ondemand.goldcast.io/on-demand/fd1ca72d-936d-49e…
Context Engineering: events.langchain.com/on-demand/64cd2752-8cf2-415f-8e64-79… -
LangChain
@LangChain
You're invited to learn how to deploy a production-ready agent in one evening using Managed Deep Agents.
🗓️ Nov 17th | 5:30-8:30
📍LangChain SF Office
🎟️ Info + RSVP: fieldevents.langchain.com/buildanddeployanagentwithmanag -
LangChain reposted
Viv
@Vtrivedy10
a couple reasons to use the deepagents harness & SDK
1. open source, extensible way to build a fully tailored agent for your use case
2. the meme game 🔥
Jess Weng
@jess_weng18
More @LangChain Deep Agents workshop memes
(Retrieval & Context Engineering edition)
Retrieval: goldcast.ondemand.goldcast.io/on-demand/fd1ca72d-936d-49e…
Context Engineering: events.langchain.com/on-demand/64cd2752-8cf2-415f-8e64-79… -
LangChain reposted
Jess Weng
@jess_weng18
More @LangChain Deep Agents workshop memes
(Retrieval & Context Engineering edition)
Retrieval: goldcast.ondemand.goldcast.io/on-demand/fd1ca72d-936d-49e…
Context Engineering: events.langchain.com/on-demand/64cd2752-8cf2-415f-8e64-79… -
LangChain
@LangChain
Headed to AWS re:Invent? On December 2nd, we're teaming up with @mongodb for a happy hour at Sugarcane!
Interested in attending? RSVP today! luma.com/ffbk3mnc -
LangChain
@LangChain
Every session, on demand.
Experience the best of Interrupt SF, NYC, and LDN: interrupt.langchain.com/recordings
LangChain
@LangChain
At Interrupt NYC, Director of ML at @Airbnb Pedro Rodriguez took the stage to share how his team built its trust and safety agents on LangGraph.
Watch the full session: www.youtube.com/watch?v=NUgaWmws0XQThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
How do you improve your agents? @divyasan0 from the LangChain Academy team walks through how the three steps of eval-driven development, a strategy for using evals to help build agents that improve over time. www.langchain.com/blog/what-is-eval-driven-development?
-
LangChain
@LangChain
At Interrupt NYC, Director of ML at @Airbnb Pedro Rodriguez took the stage to share how his team built its trust and safety agents on LangGraph.
Watch the full session: www.youtube.com/watch?v=NUgaWmws0XQThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
How do agents remember the right information at the right time?
Join Education Engineer @jess_weng18 on Monday, November 9th, for a live workshop on agent memory retrieval to find out!
Save your spot: events.langchain.com/register/agent-memory-102-context-re… -
LangChain reposted
Caspar
@caspar_br
Managed Deep Agents can now be steered mid-turn. If you send a message while the agent is working, it reads it at its next model call instead of waiting for the turn to finish. Small change that makes agents much, much nicer to work with
-
LangChain
@LangChain
Essential reading on proactive agents ⤵️
Sydney Runkle
@sydneyrunkle
we've been focused on supporting more autonomous, proactive agents; here's what we shipped in deepagents this week:
1. browser use with @Stagehanddev, so your agent can work in a real browser, sometimes on your behalf!
* blog: www.langchain.com/blog/browser-use-in-managed-deep-agents
* deep dive: www.langchain.com/blog/how-we-optimized-browser-use-for-d…
* docs: docs.langchain.com/oss/python/deepagents/browser-use
2. async subagents, so the main agent remains available to a user while background tasks run
* blog: www.langchain.com/blog/agents-are-going-async
* docs: docs.langchain.com/oss/python/deepagents/async-subagents -
LangChain
@LangChain
Agents can't hold everything in context, so they retrieve what each task needs.
This can go wrong in a few ways:
✅ Clumsy searches
✅ Oversized results
✅ Stale text
@jess_weng18 on how to combat these challenges + what to keep in mind when designing an agent harness ⤵️ -
LangChain reposted
Harrison Chase
@hwchase17
this week at langchain:
- async subagents in deep agents - your agent keeps talking to you while subagents work in the background
- how we tuned browser use in MDA (7 toolsets tested, stagehand won)
- LangSmith BYOC on AWS, Azure + GCP
- Interrupt is back, March 24-25 in SF -
LangChain
@LangChain
In SF for AI Engineer Code Summit next week? Stop by our new office for an evening of food, drinks, music, and fun as we bring the LangChain community together.
RSVP for our Friends of LangChain Party: luma.com/i4xu7a6l -
LangChain
@LangChain
Traveling in from EMEA for re:Invent?
Join us for dinner on opening night.
No slides or pitches.
Just a few of us from LangChain + a group of leaders discussing what’s working and what isn’t.
luma.com/aebfo9wt -
LangChain reposted
Jake Broekhuizen
@jakebroekhuizen
Some great mental models on writing high-quality evals by @austinberke and the harmonic team. A pleasure to work with them and some especially useful breakdowns of:
- the tradeoffs between eval dimensions rather than optimizing for one score
- evaluating verifiable correctness rather than prescribing a single "right" answer
It also nails the importance of actually getting in and looking at the data to reward responses that build 'trust' - a quality that is often overlooked if the verification is purely left to a language model.
Austin Berke
@austinberke
Scout doesn't pretend to be human, but our evals do x.com/i/article/2118421854725586944
-
LangChain reposted
nathan drezner
@ndrezn
icymi - browser use in managed deep agents!
it is a wicked great experience -
LangChain
@LangChain
Headed to re:Invent? On December 2nd, we’re hosting a breakfast panel with @baseten, @FactoryAI, and @NVIDIA to explore what it takes for a company to truly own their intelligence.
RSVP: luma.com/3kjztmsw -
LangChain reposted
Caspar
@caspar_br
Async subagents might be the biggest improvement to chat UX this year. We ship them out of the box in MDA. @sydneyrunkle wrote a great piece on how they work
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Christian Bromann
@bromann
Ask your agent in Slack to compare three vendors. Three subagents research in parallel while it keeps answering your follow-ups.
That’s async subagents in @LangChain's Managed Deep Agents. No setup needed: every served agent gets an async general-purpose subagent.
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
We revamped support for async subagents in (Managed) Deep Agents!
Learn about how they keep your orchestrator agent responsive to the user while making progress on tasks in the background ⬇️
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
LangChain OSS
@LangChain_OSS
Async subagents keep your agent responsive to you while background workers handle the long tasks.
Try them out in Deep Agents today!
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Sydney Runkle
@sydneyrunkle
i messed up the code block formatting in my blog on "browser use for deep agents" the design agent we tested browser use in fixed it, then used browser use to validate the change.
great example of how browser use enhances testing, very full circle!
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
👀
amal
@amal_irgashev
im telling you, you gotta try content studio...
luckily for you, its open source
github.com/langchain-ai/design-agentThis post is unavailable. See it on X. -
LangChain
@LangChain
LangChain? LangGraph? Deep Agents?
The full ecosystem explained in 3 minutes.This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
We have several LangChain-led and community-led meetups coming up through the rest of the year! RSVP and bring a friend!
🌉 11/5 San Francisco
🇬🇧 11/12 London
🇺🇸 11/12 Boston
🌉 11/10 SF
🌉 11/12 SF
🗽 11/17 NY
🇺🇸 11/18 Seattle
🇵🇱 11/24 Kraków
🇸🇪 11/25 Stockholm
🇩🇪 12/4 Hamburg
RSVP 👉 luma.com/langchain -
LangChain reposted
Sydney Runkle
@sydneyrunkle
full transparency, things that went wrong during the filming of this demo...
- almost booked a reservation that required a HUGE downpayment / cancellation fee bc my card was saved on resy, so the agent could just book+pay
- hit rate limits w/ resy bc i simulated log in so many times trying to get the demo right
browser use is a relatively new primitive for agents, so it makes sense that we're working through some growing pains
the hardest things rn seem to be guardrails + working at scale
Sydney Runkle
@sydneyrunkle
if you give your agent a browser, it can do work on the web, even on sites w/ no agent native APIs!
in this demo my agent books dinner reservations for me. i only had to log into the reservation site once; it saved my login and acts on my behalf for every new session!This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Watch @sydneyrunkle use an agent to book dinner reservations, powered by browser use in Managed Deep Agents!
Sydney Runkle
@sydneyrunkle
if you give your agent a browser, it can do work on the web, even on sites w/ no agent native APIs!
in this demo my agent books dinner reservations for me. i only had to log into the reservation site once; it saved my login and acts on my behalf for every new session!This media couldn't be displayed. See it on the original post. -
LangChain reposted
Caspar
@caspar_br
A year ago I wrote that LangGraph isn't going anywhere. Today it's the runtime under everything at LangChain, including MDA
@JoshARosen wrote a good piece on why deterministic layers will become standard agent infrastructure. We made the same bet early: x.com/JoshARosen/status/2117687393213932003
Caspar
@caspar_br
The fact that LangGraph has not changed much is a testament to its robustness and timelessness. I've loved working on this framework and I'm continuously blown away by the diversity of agentic applications I've seen built with it.
This is the decade of agents and LangGraph isn't going anywhere. Check out our new docs site which unifies langchain and langgraph OSS docs, there are some great tutorials to get you started. Knowing how to build with this framework is an insanely useful skillThis post is unavailable. See it on X. -
LangChain reposted
Tavily
@tavilyai
Who's joining us for World of Tavily NYC? On November 12, you'll hear lightning talks from leaders at Tavily, @nebiusai, @LangChain, Ocean, and @IBM about how production agents find, understand, and use information from the live web.
Come ready to ask questions, exchange ideas with other builders, and spend time with the speakers. We’ll also have plenty of bites, drinks, swag, and immersive sights to explore.
📅 Thursday, November 12
📍 New York City
Request to join: luma.com/ivrwjk0o?hss_channel=tw-1716430063191621633 -
LangChain
@LangChain
Headed to re:Invent?
We’re teaming up with @traversal, @factory, @sequoia, @p0, and @temporal for an evening of Michelin-caliber dining, great drinks, and conversation about the future of AI.
RSVP while spaces last!
luma.com/agents-table-at-reinvent -
LangChain reposted
Sydney Runkle
@sydneyrunkle
this piece is worth the read, but quite dense. here's my (relatively) brief breakdown:
super agents are defined by 4 characteristics:
- continuous: one agent across threads and access points
- productive: responds quickly and completes work efficiently
- proactive: responds to events in its environment
- persistent: durable execution for high availability and crash recovery
these constraints apply to personal agents and agents doing commercial work alike, so the underlying architecture should look similar
the primitives behind them:
- two level architecture: a router layer keeps interaction continuous, and a worker layer runs tasks concurrently
- event driven design: the agent reacts to events in its environment and can schedule runs toward a long term goal
- multi channel interaction: context carries across channels, and async execution keeps the agent productive
- separate compute, state, and environments: the "separate the brain from the hands" idea, which supports persistence and security and lets the system scale
- watchdog: an independent control loop that keeps the whole thing running
they're upfront about what's still open:
- reliable scale
- representing tasks and work without the "session" primitive
- agent to agent communication and context standards
- security and privacy, as always
their core claim is that models need less and less human intervention on individual tasks, so the bottleneck to further productivity gains is shifting to human attention
which raises the question i'm most interested in: how do we design human-agent collaboration when human attention is the bottleneck?
DeepSeek Harness
@DeepSeekHarness
-
LangChain reposted
Sydney Runkle
@sydneyrunkle
there is a TON of value in effective tool design
we tested 7 different toolsets against a relatively saturated browser use benchmark to figure out which toolsets were token/cost efficient
@Stagehanddev from @browserbase was the clear winner here!
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Ankush Gola
@ankush_gola11
Come to our @ApacheDataFusio meetup next Tuesday at our SF office!
LangChain
@LangChain
📍 Happening November 10th: @ApacheDataFusio SF meetup
Join us for an evening of talks and community discussion about Apache DataFusion and its growing role in modern data infrastructure.
Spaces are limited! RSVP today: luma.com/k04h0om1 -
LangChain
@LangChain
📍 Happening November 10th: @ApacheDataFusio SF meetup
Join us for an evening of talks and community discussion about Apache DataFusion and its growing role in modern data infrastructure.
Spaces are limited! RSVP today: luma.com/k04h0om1 -
LangChain reposted
Sydney Runkle
@sydneyrunkle
we're hosting an event on thursday at the @LangChain office with @allietheicon from @typesafeai to chat about all things jev!
come learn about
- where jev fits in the agent loop
- why jev is a "system one" model
- jev as a judge for evals
luma.com/iuc3ostf -
LangChain
@LangChain
LangSmith Bring Your Own Cloud (BYOC) now runs on @AWScloud, @Azure, + @GoogleCloud.
Your traces and agent deployments stay in your cloud account while we run the data plane.
☁️ GA on AWS, Beta on Azure + Google Cloud
🚚 Moving from self-hosted? We'll run the migration -
LangChain
@LangChain
Proud to be a launch partner in the @FireworksAI_HQ Frontier Ecosystem!
Fireworks
@FireworksAI_HQ
The future of AI is specialized intelligence with models tuned to your data, your domain and your customers.
Today, we're proud to launch live from Forge the Fireworks Frontier Ecosystem with 60+ partners across the AI stack, helping you build and ship faster.
Check it out: -
LangChain
@LangChain
Browser use lets agents act on sites built for people, and it puts real pressure on the context window.
Here's how we chose and tuned the browser tools in Managed Deep Agents, built on @Stagehanddev from @browserbase
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
The complete LangChain ecosystem explained in three minutes.
This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
ICYMI
Harrison Chase
@hwchase17
A lot of content engineering challenges have pretty standard solutions, which has lead to some engineers saying agent engineering is solved/boring
True to some extent, but there are a bunch of (fun) new ones emerging as agents evolve! x.com/i/article/2116547440857718784 -
LangChain reposted
Viv
@Vtrivedy10
gonna be another banger with awesome ppl 👀
come hang & we can yap about Evals, Swarms, Post-Training & other fun stuff :)
Harrison Chase
@hwchase17
We're bringing back Interrupt for the THIRD year
- This will be the biggest one yet (new venue that can hold ~2k people!)
- Most talks & workshops we've ever had
- New Executive Track
Register today for the early bird special: interrupt.langchain.com -
LangChain reposted
Sydney Runkle
@sydneyrunkle
agentic browser requests jumped ~5x in september alone... browser use is becoming a key primitive for agents!
icymi, we shipped it in (managed) deep agents last week
now your agent can do anything interactive on the web!
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
amal
@amal_irgashev
we are so back
LangChain
@LangChain
🌉 Interrupt, The Agent Conference by LangChain, returns to San Francisco March 24 + 25
🎟️ Early Bird tickets available until Dec 7th
Info + RSVP: interrupt.langchain.com -
LangChain reposted
Harrison Chase
@hwchase17
We're bringing back Interrupt for the THIRD year
- This will be the biggest one yet (new venue that can hold ~2k people!)
- Most talks & workshops we've ever had
- New Executive Track
Register today for the early bird special: interrupt.langchain.com -
LangChain
@LangChain
Interested in a career with LangChain?
On November 12th, come to our NYC Office to meet the team and kickstart your sales career in a Sales Development Representative (SDR) role.
RSVP ⤵️ luma.com/langchain-np4d -
LangChain reposted
Sydney Runkle
@sydneyrunkle
we think a lot about context engineering at langchain
there's a lot of of old but good primitives here: skills, tool search, summarization to name a few
this short blog covers new challenges in this space: browser use, long term goal tracking, async work, and more!
Harrison Chase
@hwchase17
A lot of content engineering challenges have pretty standard solutions, which has lead to some engineers saying agent engineering is solved/boring
True to some extent, but there are a bunch of (fun) new ones emerging as agents evolve! x.com/i/article/2116547440857718784 -
LangChain
@LangChain
🌉 Interrupt, The Agent Conference by LangChain, returns to San Francisco March 24 + 25
🎟️ Early Bird tickets available until Dec 7th
Info + RSVP: interrupt.langchain.com -
LangChain
@LangChain
How do agents remember?
And what does preventing context rot have to do with it?
Join our live workshop on context engineering for agent memory and learn:
✅ Context management techniques by building w/ Deep Agents
✅ How to keep the context window lean from the first tool call to final report
events.langchain.com/register/agent-memory-101-context-en… -
LangChain reposted
Harrison Chase
@hwchase17
A lot of content engineering challenges have pretty standard solutions, which has lead to some engineers saying agent engineering is solved/boring
True to some extent, but there are a bunch of (fun) new ones emerging as agents evolve! x.com/i/article/2116547440857718784 -
LangChain reposted
Harrison Chase
@hwchase17
want to learn how to use decision models with langgraph/deepagents?
we're doing a meetup with Typesafe (Jev creators) next week
come join! luma.com/iuc3ostf -
LangChain
@LangChain
LangChain Community Spotlight: Enterprise RAG on LangChain 🔎📚
Kacper Klusek of Soba Labs shows how to build enterprise RAG on LangChain by fixing retrieval before you touch the model. In Anthropic's published evaluation, the retrieval chain he recommends (contextual embeddings, BM25 and a reranker) cut failed retrievals by 67%. LangSmith online evals then catch the decay that starts when your documents change.
📖 Read the full guide: sobalabs.ai/blog/enterprise-rag-on-langchain/ -
LangChain reposted
nathan drezner
@ndrezn
nearly every proactive agent you've seen is a cron job with a prompt
memory that "dreams" overnight, agents that ping you first. both are scheduled runs on a 50+ year old OS idea
how to build one with Managed Deep Agents ⬇️ x.com/i/article/2115474237418778624 -
LangChain
@LangChain
There's still time to RSVP for our upcoming SF meetup with @typesafeai
Join @sydneyrunkle, @allietheicon, and Seán Roche at LangChain HQ for a night of talks on building with decision models and using them to evaluate agents.
Spaces are limited!
luma.com/iuc3ostf -
LangChain
@LangChain
If you need a framework for thinking about the loops around an agent and want to learn how to build them, look no further.
Learn about Loop Engineering with Deep Agents today: www.deeplearning.ai/courses/loop-engineering-with-langcha…
DeepLearning.AI
@DeepLearningAI
Build reliable agents for long-running tasks.
Loop Engineering with @LangChain Deep Agents: add loops that check output, trigger runs on events, and improve the harness from traces across runs. Taught by @sydneyrunkle and @EAGar123.
Enroll for free: www.deeplearning.ai/courses/loop-engineering-with-langcha… -
LangChain
@LangChain
.@hwchase17 kicked off Interrupt London with a keynote on where agents are headed, and why every org should own their intelligence.
🔴 Full talk: youtu.be/XOUSuxY074U
⏯️ Every session, on demand: interrupt.langchain.com/recordingsThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
LangSmith for Startups Spotlight: @PetualAI
✅ Agents for internal audit and SOX teams at leading public companies and high-growth enterprises.
✅ Use LangSmith within their engineering workflow to understand, improve, and validate the performance of their AI systems.
🏢 Hiring across functions: petual.ai/careers -
LangChain
@LangChain
“Typically the manual mechanism takes about 20 days. Just with the first pass, we are able to drive it down to 10.”
@skyworksinc uses LangGraph and LangSmith to deploy agents and run evals in one place.
⏯️ How they’re using agents to cut root cause analysis time.This media couldn't be displayed. See it on the original post. -
LangChain reposted
Harrison Chase
@hwchase17
a lot of the challenges in harness engineering are now best solved by closer integration of harness and infra
async subagents and browser use are two examples of this
now in MDA 0.10!
nathan drezner
@ndrezn
most of the web has no API, just pages built for people clicking through them
so! we added a browser to Managed Deep Agents. and when a site wants a login, agents can hand over the window so passwords stay out of the model.
all built in with every MDA x.com/i/article/2115851731741347840 -
LangChain
@LangChain
Managed Deep Agents 0.10: Now available in public beta
🌐 Browser use: Your agent can now work on websites with no API
💬 Async subagents: Run long tasks in the background while your agent keeps talking
🧵 Thread origins and stopping runs
...and so much more
nathan drezner
@ndrezn
most of the web has no API, just pages built for people clicking through them
so! we added a browser to Managed Deep Agents. and when a site wants a login, agents can hand over the window so passwords stay out of the model.
all built in with every MDA x.com/i/article/2115851731741347840 -
LangChain reposted
amal
@amal_irgashev
a lot of you asked how i made our launch video for Content Studio
it took 13 iterations, 5 hours and ~$120, all done in a Slack thread
here's the breakdown of every step 🧵
amal
@amal_irgashev
We just open sourced our internal design brain 🧠
Meet Content Studio - the agent behind 3,000+ assets at @LangChain
Create on-brand product launch videos, slide decks, ad creatives, social media assets, ui mockups and more
Storyboard and collaborate on a shared canvas in @paper and edit videos in code with @HyperFrames_
It even made its own launch film below, and went full Rocky 🥊
Fork it, add your brand guidelines, and deploy it to your team in Slack with Managed Deep Agents
Here’s what 's inside 🧵 -
LangChain
@LangChain
Excited to partner with the @Eon_io_ team to benchmark model misalignment in enterprise environments.
The first question we're digging into: Do results change when the model runs inside the Deep Agents harness? Stay tuned for more research.
Ofir Ehrlich
@OfirEhrlich
We just ran the most comprehensive benchmark of AI agent alignment in enterprise environments:
19 leading AI models acting as agents inside simulated companies built with Era, 2,304 runs across 41 everyday tasks.
Agents broke security and business rules in 40% of those runs. Every single model failed, including all the frontier models. A few examples:
- An agent told one of the company's biggest customers that Finance had approved a full refund.
Nobody had.
- An agent shared a customer's internal churn-risk score directly with that customer.
- An agent followed a fake runbook planted in a support ticket and promised a customer a credit nobody approved.
- An agent posted another customer's personal contact details (PII) in a shared Slack channel.
In a real company, each of these is a financial liability, a compliance breach, or a lost account.
We've shared the full results with the model providers, including exploitable failure modes we're not publishing, and are working with them on fixes.
One benchmark won't fix this.
We're excited to have @cognition, @LangChain, @TryArcade, @Reindeer_AI, Hirundo, @Box, @recolabs_ai and @QodoAI join Era as research partners, as we build out the next round of research.
More soon.
I'm still the most bullish person in the room on AI.
But this benchmark convinced me that agent reliability and trust is the next big problem to solve.
Until then, test every agent before it gets a login, and keep your own record of what it changes.
Full results: era.eon.io/model-alignment-benchmark.html -
LangChain
@LangChain
🌐 Managed Deep Agents can now use a browser
Your agent can:
- Take actions on sites with no API, like booking a reservation
- Test an application end to end through its UI
- Hand you the browser to sign in, then pick up where it left off
Docs ➡️ docs.langchain.com/langsmith/python/managed-deep-agents-b…
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
amal
@amal_irgashev
one of the biggest things i learned building @LangChain’s internal design agent is that most of the rules i wrote actually made it worse...
they were there to compensate for the model’s intelligence at the time
but when Opus 5.5 came out i deleted most of them, and the work got a lot more creative
we got into this with 350+ people at our GTM Engineering Live today.
key takeaways below 🧵 -
LangChain
@LangChain
LangChain is returning to re:Invent. A TL;DR:
📍 Meet us at Booth #1740
🎙️ Talks + Panels throughout the week
🗓️ Co-hosted events featuring breakfasts, networking, dinners, and more
See what we’re up to + book meetings: fieldevents.langchain.com/awsreinvent2026 -
LangChain reposted
Colin Francis
@colifran_
we've been exploring how to make the knowledge openwiki produces more useful to agents.
our latest experiment pairs the wiki with a lightweight evidence graph, letting agents follow relationships and trace information back to supporting evidence.
on complex, domain-specific questions requiring cross-source reasoning we saw:
👉 5x higher accuracy with openwiki vs. grep alone
👉 27% more accurate with the evidence graph
👉 2x citation consistency
👉 58% lower cost per correct answer with the graph
check out the blog ⤵️
www.langchain.com/blog/building-agent-memory-openwiki-graphs -
LangChain
@LangChain
Great look at making agent memory traceable to evidence ⤵️
Colin Francis
@colifran_
we've been exploring how to make the knowledge openwiki produces more useful to agents.
our latest experiment pairs the wiki with a lightweight evidence graph, letting agents follow relationships and trace information back to supporting evidence.
on complex, domain-specific questions requiring cross-source reasoning we saw:
👉 5x higher accuracy with openwiki vs. grep alone
👉 27% more accurate with the evidence graph
👉 2x citation consistency
👉 58% lower cost per correct answer with the graph
check out the blog ⤵️
www.langchain.com/blog/building-agent-memory-openwiki-graphs -
LangChain reposted
Sydney Runkle
@sydneyrunkle
"the art of loop engineering" is now a @DeepLearningAI course!
www.deeplearning.ai/courses/loop-engineering-with-langcha…
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
amal
@amal_irgashev
deslopify - made with Content Studio
This media couldn't be displayed. See it on the original post.
amal
@amal_irgashev
We just open sourced our internal design brain 🧠
Meet Content Studio - the agent behind 3,000+ assets at @LangChain
Create on-brand product launch videos, slide decks, ad creatives, social media assets, ui mockups and more
Storyboard and collaborate on a shared canvas in @paper and edit videos in code with @HyperFrames_
It even made its own launch film below, and went full Rocky 🥊
Fork it, add your brand guidelines, and deploy it to your team in Slack with Managed Deep Agents
Here’s what 's inside 🧵 -
LangChain reposted
DeepLearning.AI
@DeepLearningAI
Build reliable agents for long-running tasks.
Loop Engineering with @LangChain Deep Agents: add loops that check output, trigger runs on events, and improve the harness from traces across runs. Taught by @sydneyrunkle and @EAGar123.
Enroll for free: www.deeplearning.ai/courses/loop-engineering-with-langcha… -
LangChain reposted
Sydney Runkle
@sydneyrunkle
great to work w/ @AndrewYNg and @EAGar123 on this loop engineering course!
this course focuses on how to build agents that can a) handle complex tasks and b) improve over time with deepagents!
DeepLearning.AI
@DeepLearningAI
Build reliable agents for long-running tasks.
Loop Engineering with @LangChain Deep Agents: add loops that check output, trigger runs on events, and improve the harness from traces across runs. Taught by @sydneyrunkle and @EAGar123.
Enroll for free: www.deeplearning.ai/courses/loop-engineering-with-langcha… -
LangChain
@LangChain
You’re an mda init and mda deploy away from distributing a proactive agent in your org.
Since launching dynamic scheduling for Managed Deep Agents in September, over 70% of MDA runs on LangSmith come from schedules vs human typing a message.
nathan drezner
@ndrezn
nearly every proactive agent you've seen is a cron job with a prompt
memory that "dreams" overnight, agents that ping you first. both are scheduled runs on a 50+ year old OS idea
how to build one with Managed Deep Agents ⬇️ x.com/i/article/2115474237418778624 -
LangChain
@LangChain
The company? The framework?
LangChain explained in 2 minutes.This media couldn't be displayed. See it on the original post. -
LangChain reposted
Caspar
@caspar_br
Company harnesses break down knowledge silos
Three in four requests to Content Studio come from people outside marketing. This usage pattern is why I'm so bullish on company harnesses: they allow everyone to get involved in domains they had no business in before. We've encoded our design team's judgment into a set of skills anyone can use, so anyone with a creative vision can execute it with Content Studio
Coding agents brought the same shift to SWE. People who would have called themselves nontechnical a year ago are now shipping PRs
I'm excited to see this happen in every domain
LangChain
@LangChain
How we built Content Studio, an open-source design agent in Slack that makes on-brand videos, graphics and decks, and the harness that powers it. x.com/i/article/2114731014794661888
-
LangChain
@LangChain
LangSmith now tracks cache hit rates for your agents 🦜
Cached LLM calls are cheaper and faster, so it's worth knowing how often your agent actually gets a hit.
📊 View your agent's overall hit rate as a single stat
🧵 Break it down thread by thread
📈 Chart it over time with our dashboard templates
Try it out today ➡️ smith.langchain.com -
LangChain
@LangChain
Our design agent doesn’t burn tokens starting over. It never hands back a flat image or a generated clip.
Slides are HTML. Charts are data specs. Films are animation timelines. “Move the parrot left” is a one-value change, not a regenerate-from-scratch.
🍿 Want to go deeper? We asked our design agent to make a film about how it works.This media couldn't be displayed. See it on the original post.
LangChain
@LangChain
How we built Content Studio, an open-source design agent in Slack that makes on-brand videos, graphics and decks, and the harness that powers it. x.com/i/article/2114731014794661888
-
LangChain
@LangChain
Come learn about our design agent on October 28th!
Chat with our team, see it in action, and discover how it works!
RSVP: events.langchain.com/gtm-engineering-live-series/design-a…This media couldn't be displayed. See it on the original post.
LangChain
@LangChain
How we built Content Studio, an open-source design agent in Slack that makes on-brand videos, graphics and decks, and the harness that powers it. x.com/i/article/2114731014794661888
-
LangChain
@LangChain
Built on Managed Deep Agents. Used by our team every day.
LangChain
@LangChain
How we built Content Studio, an open-source design agent in Slack that makes on-brand videos, graphics and decks, and the harness that powers it. x.com/i/article/2114731014794661888
-
LangChain reposted
Harrison Chase
@hwchase17
deepdive into how we built our design agent (along with oss code)
i use this at least a few times a week for slides, gifs, images, etc
LangChain
@LangChain
How we built Content Studio, an open-source design agent in Slack that makes on-brand videos, graphics and decks, and the harness that powers it. x.com/i/article/2114731014794661888
-
LangChain reposted
nathan drezner
@ndrezn
bet ya didn't know... Content Studio is a managed deepagent!!!
want your own content studio deployed in your organization?
git clone
mda deploy
boom
LangChain
@LangChain
How we built Content Studio, an open-source design agent in Slack that makes on-brand videos, graphics and decks, and the harness that powers it. x.com/i/article/2114731014794661888
-
LangChain reposted
amal
@amal_irgashev
We just open sourced our internal design brain 🧠
Meet Content Studio - the agent behind 3,000+ assets at @LangChain
Create on-brand product launch videos, slide decks, ad creatives, social media assets, ui mockups and more
Storyboard and collaborate on a shared canvas in @paper and edit videos in code with @HyperFrames_
It even made its own launch film below, and went full Rocky 🥊
Fork it, add your brand guidelines, and deploy it to your team in Slack with Managed Deep Agents
Here’s what 's inside 🧵 -
LangChain
@LangChain
How we built Content Studio, an open-source design agent in Slack that makes on-brand videos, graphics and decks, and the harness that powers it. x.com/i/article/2114731014794661888
-
LangChain
@LangChain
.@Box’s Deep Agent architecture for Box Agent uses a parent/child model.
Simple request? The parent handles it directly. No planning needed.
Complex request? It spawns child agents to distribute the work, keeping the invocation surface uniform throughout.
LangChain
@LangChain
"If the question was very simple, or the search was very simple, the parent node can just do it. It doesn’t even need to come up with a plan”
-Shubhro Roy, AI Engineering Leader at @Box
www.langchain.com/blog/building-box-ai-how-an-enterprise-… -
LangChain
@LangChain
What does it take to get to an agent that catches its own mistake and writes the eval and the fix?
Catch @hwchase17's Interrupt London session on building agents that improve over time: youtu.be/xy_7SzwsIjM -
LangChain
@LangChain
Experience the best of Interrupt 26.
📍 SF
📍 NYC
📍 LDN
Every session. On demand. interrupt.langchain.com/recordings
LangChain
@LangChain
What’s new in Managed Deep Agents 0.9? And how can you get started?
Watch @VictorMoreira16's keynote from Interrupt London for an inside look. youtu.be/Wb3hQaG_xf0This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Can you trust an LLM judge?
Join @divyasan0 Oct 28th for a virtual workshop on:
✅ How to score a trajectory with the trajectory match evaluator
✅ How to tune matching so evals fail on real behavior changes, not argument formatting
✅ When to use deterministic matching vs. LLM-as-judge
✅ How to wire evals into your loop so regressions show up before they ship
Interested in attending? RSVP today! events.langchain.com/register/762756f4-4344-4760-b4b9-a07… -
LangChain
@LangChain
Join LangChain + @typesafeAI on November 5th for an evening of talks on building with decision models and using them to evaluate agents.
Hear from @sydneyrunkle, Seán Roche, and @allietheicon.
Interested in attending? RSVP today! luma.com/iuc3ostf -
LangChain
@LangChain
LangGraph explained in 1 minute.
This media couldn't be displayed. See it on the original post. -
LangChain reposted
amal
@amal_irgashev
We learned a ton building the internal design brain for the company from building a custom media harness to encoding taste and design judgement.
Will share all learnings in Wednesday!
LangChain
@LangChain
💻 Happening this Wednesday!
Go behind the scenes of the design agent we use every day at LangChain.
RSVP for our GTM Engineering live series event:
events.langchain.com/gtm-engineering-live-series/design-a… -
LangChain
@LangChain
💻 Happening this Wednesday!
Go behind the scenes of the design agent we use every day at LangChain.
RSVP for our GTM Engineering live series event:
events.langchain.com/gtm-engineering-live-series/design-a… -
LangChain reposted
Sydney Runkle
@sydneyrunkle
giving a talk tomorrow on "building agents that work like coding agents"
at @LangChain we help folks build all types of agents, but a lot of them "work like" coding agents in terms of harness design
luma.com/cico2dxy -
LangChain
@LangChain
TL;DR on Managed Deep Agents 0.9:
✅ Agent-created schedules: agents can set up reminders and recurring tasks from a conversation.
✅ Per-run configuration: one deployment can pick its model, skills, and tools for each run.
✅ Slack reactions: agents acknowledge messages before they reply.
www.langchain.com/blog/managed-deep-agents-schedules-per-…
LangChain
@LangChain
What’s new in Managed Deep Agents 0.9? And how can you get started?
Watch @VictorMoreira16's keynote from Interrupt London for an inside look. youtu.be/Wb3hQaG_xf0This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
What’s new in Managed Deep Agents 0.9? And how can you get started?
Watch @VictorMoreira16's keynote from Interrupt London for an inside look. youtu.be/Wb3hQaG_xf0This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
On November 12th, we're teaming up with @tavilyai to showcase how production agents find, understand, and use information from the live web.
RSVP for this NYC event: luma.com/ivrwjk0o -
LangChain
@LangChain
Continual learning is about to get boring, in a good way.
@willcb from @primeintellect took the stage at our recent Learning Loop event to make the case.
If you're thinking about how continual learning fits into the way your team already ships software, this full 21-minute session is a good place to start.This media couldn't be displayed. See it on the original post. -
LangChain reposted
amal
@amal_irgashev
taste is the judgment that separates work that's correct from work that's good
models can already write html, css, chart specs and animation code. what they don't have is your context, so they reach for the same familiar patterns (i.e distributional convergence)
our design agent at @LangChain did every ai slop pattern imaginable: edge-bar cards, a label above every title, words · like · this, and dot-grid wallpaper behind everything
so we updated the system so that anything measurable is a check in code and judgment lives in skills and a design wiki. design principles like one focal point per frame, a guardrail for each slop pattern, and the reasoning behind every note our designers give
anything someone teaches it in slack gets saved to memory in the same turn
we also tried prebuilding taste with 115 UI components so the model had fewer decisions to make but it just it kept repeating the same layouts and animations...
so when opus 5.5 was released, we actually deleted most of those rules and kept the principles. more variety, way less repetition
we are going deeper on all of it live oct 28 in our GTM Engineering series with a special guest, our goated brand designer, @shreyatgkr
register here 👇
events.langchain.com/gtm-engineering-live-series/design-a… -
LangChain
@LangChain
We're hiring!
Check out our open roles @ www.langchain.com/careers
amal
@amal_irgashev
i think the best part of building internal agents is that you get to sit with people who've done this stuff for years and just... absorb
at @LangChain i get to work with some of the best GTM people i've ever met. people who've run huge launches, built massive brands, closed hundreds of deals and know exactly what great looks like
a big part of the job is basically an apprenticeship. you shadow them, interview them, figure out how they make the calls they make, and translate that judgment into context agents can use - an insanely fast feedback loop
and it won't work right away. at first it's this slightly alien thing in slack they have to babysit, and it can take hundreds of iterations before people really start trusting it
but then one day it just clicks. someone asks for something new and it comes back exactly how they would've done it, with zero notes. from then on it's just part of their day, and you get to watch it change the way they work
and the most fascinating part is that the agent ends up being the best written record of how the business actually works
honestly one of the most rewarding jobs i've had -
LangChain
@LangChain
🎓 New from the LangChain Academy team
How agents remember to retrieve the right info at the right time ⤵️
Jess Weng
@jess_weng18
I present to you 🥁🥁🥁
Context Engineering: The Prequel (Retrieval)
www.langchain.com/blog/context-engineering-starts-with-re… -
LangChain
@LangChain
The best of Interrupt, now on demand.
Access our archive today and start watching our sessions from Interrupt San Francisco, NYC, and London. interrupt.langchain.com/recordings -
LangChain
@LangChain
Next stop on the LangSmith Roadshow? Chicago!
Hear about the current state of agent development, and get ready for two hands-on workshops focused on Deep Agents and LangSmith Engine.
Interested in attending? RSVP today!
events.langchain.com/LangSmithRoadshow/Chicago/ -
LangChain
@LangChain
We have several LangChain-led and community-led meetups coming up! RSVP and bring a friend!
🇩🇪 10/27 Munich
🇺🇸 10/29 Washington, DC
🇲🇽 10/29 Mexico City
🇦🇪 10/31 Dubai
🇨🇦 11/4 Toronto
🇺🇸 11/5 San Francisco
🇬🇧 11/12 London
🇺🇸 11/12 Boston
🇺🇸 11/18 Seattle
🇸🇪 11/25 Stockholm
🇩🇪 12/4 Hamburg
RSVP: luma.com/langchain -
LangChain
@LangChain
Join us, @ExtendHQ, and @tryramp this Thursday for a private screening of You Can See Everything in San Francisco.
Seats are limited, save yours now! luma.com/27b59c8i -
LangChain reposted
Caspar
@caspar_br
Content Studio is a truly epic example of a company harness built on Managed Deep Agents. It is multiplayer by nature, I work with it in a public channel so others can weigh in on the design process
amal
@amal_irgashev
our design agent at @LangChain has shipped 3,000+ assets since august, and posts with its work pulled ~4.5M views on socials (it made this promo too)
it’s become the design brain for the whole org. decks, diagrams, ad creatives, launch films and more, all made from slack
we encoded our designers’ judgment into its skills, a design wiki and brand checks that run on every render, so it knows what good looks like for us
then we built a custom media harness around it: a headless photoshop, a shared visual canvas (thanks @paper), video tools like @HyperFrames_ plus our own animation engine, our brand assets and a library of design references
it runs on Managed Deep Agents, which abstracts away a ton of infra including sandboxes, agent memory, slack integration, scheduling and more.
we are going through the architecture, use cases and impact live at our oct 28 webinar 👇
events.langchain.com/gtm-engineering-live-series/design-a…This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
On October 28th, go behind the scenes of our design agent.
We’ll demo the agent live, walk through how we built it, and share what we’ve learned from using an agent for creative production across the company.
💻 RSVP: events.langchain.com/gtm-engineering-live-series/design-a…
amal
@amal_irgashev
our design agent at @LangChain has shipped 3,000+ assets since august, and posts with its work pulled ~4.5M views on socials (it made this promo too)
it’s become the design brain for the whole org. decks, diagrams, ad creatives, launch films and more, all made from slack
we encoded our designers’ judgment into its skills, a design wiki and brand checks that run on every render, so it knows what good looks like for us
then we built a custom media harness around it: a headless photoshop, a shared visual canvas (thanks @paper), video tools like @HyperFrames_ plus our own animation engine, our brand assets and a library of design references
it runs on Managed Deep Agents, which abstracts away a ton of infra including sandboxes, agent memory, slack integration, scheduling and more.
we are going through the architecture, use cases and impact live at our oct 28 webinar 👇
events.langchain.com/gtm-engineering-live-series/design-a…This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Want to learn more about agent memory?
We're hosting a virtual workshop on November 3rd! events.langchain.com/register/agent-memory-101-context-en…
LangChain
@LangChain
Context Engineering explained in 5 minutes by @jess_weng18.
This media couldn't be displayed. See it on the original post. -
LangChain reposted
Ankush Gola
@ankush_gola11
amazing to see such a successful example of using agent traces in LangSmith to fine tune a model! Was awesome working with the @corridor team on this
LangChain
@LangChain
.@corridor used smithtune’s trace-based workflow to build the training data for Argos, a fine-tuned 9B model that scores the highest of the 14 models they tested on their security context benchmark.
🧵 Full thread ⤵️
Corridor
@corridor
Introducing corridor-argos-1, Corridor’s first security model.
A Qwen 3.5 9B fine-tune that outperforms frontier models including GPT-6 Astra and Claude Opus 5.5 on security context selection, with the lowest time to first token of the 14 models we tested.
Thread 🧵 -
LangChain reposted
Jake Broekhuizen
@jakebroekhuizen
Hats off to @Aditi and the team at @corridor for the release of Argos 1! 🎩
This is an awesome example of agent trajectories in LangSmith -> post-trained model in practice, with 𝚜𝚖𝚒𝚝𝚑𝚝𝚞𝚗𝚎 helping make it possible.
They sft'd qwen3.5-9B and ended up with a model that beats gpt-6, opus 5.5, and gemini 3.6 flash on both quality and response time when selecting relevant security guidance for coding agents.
This kind of process is a great way to build alpha into your agent, and we're super bullish on giving teams the ability to use their own data to drive improvements at the model layer.
aditi nara 🍎
@aditinaraa
@farzaankai and I are so excited to share what we've been working on over the last few months.
Introducing Argos: Corridor's first security model.
Special thanks to @baseten and @LangChain for all your support.
Corridor
@corridor
Introducing corridor-argos-1, Corridor’s first security model.
A Qwen 3.5 9B fine-tune that outperforms frontier models including GPT-6 Astra and Claude Opus 5.5 on security context selection, with the lowest time to first token of the 14 models we tested.
Thread 🧵 -
LangChain
@LangChain
.@corridor used smithtune’s trace-based workflow to build the training data for Argos, a fine-tuned 9B model that scores the highest of the 14 models they tested on their security context benchmark.
🧵 Full thread ⤵️
Corridor
@corridor
Introducing corridor-argos-1, Corridor’s first security model.
A Qwen 3.5 9B fine-tune that outperforms frontier models including GPT-6 Astra and Claude Opus 5.5 on security context selection, with the lowest time to first token of the 14 models we tested.
Thread 🧵 -
LangChain
@LangChain
Context Engineering explained in 5 minutes by @jess_weng18.
This media couldn't be displayed. See it on the original post. -
LangChain reposted
-
LangChain reposted
Divya Santhanam
@divyasan0
Agent evals raise a lot of questions. We rounded up the most common ones and answered them! Take a look 👀 x.com/i/article/2112615367663341568
-
LangChain
@LangChain
Agents are being asked to take on increasingly complex work, from research and coding to workflows that span multiple tools and steps.
The challenge is giving them the structure to do that work reliably.
Join our APAC + ANZ Agent Harnesses & Deep Agents 101 Virtual Workshop to explore the building blocks behind agents designed for longer-horizon tasks and learn how to apply those patterns in practice.
events.langchain.com/register/apac-anz-agent-harnesses-an… -
LangChain reposted
Harrison Chase
@hwchase17
great points by caspar on what makes building a "company harness" (companyOS? org harness?) hard
and how MDA solves them!
Caspar
@caspar_br
How do you build a company harness?
We've built internal agents at LangChain across marketing, sales, and engineering. The results have been incredible but these agents were NOT easy to get right. We spent a ton of time solving the problems that come with a multiplayer harness. Every agent team landed on the same core primitives:
- It lives where people already work (for us that's Slack)
- It knows who's asking and acts OBO that person for some services and with its own identity for others
- It has memory, scoped globally and/or at the user level
- It runs code in a sandbox where authenticated CLIs and scripts just work (gh, curl, your internal APIs)
- Large % of its runs are scheduled rather than explicitly invoked by someone (ambient agents!!)
It became clear we needed one shared place for these primitives, so we built Managed Deep Agents. In my humble opinion it is the easiest way to build a company harness today
Our docs are very agent-friendly, point them here to get started: langch.in/mda -
LangChain reposted
Caspar
@caspar_br
How do you build a company harness?
We've built internal agents at LangChain across marketing, sales, and engineering. The results have been incredible but these agents were NOT easy to get right. We spent a ton of time solving the problems that come with a multiplayer harness. Every agent team landed on the same core primitives:
- It lives where people already work (for us that's Slack)
- It knows who's asking and acts OBO that person for some services and with its own identity for others
- It has memory, scoped globally and/or at the user level
- It runs code in a sandbox where authenticated CLIs and scripts just work (gh, curl, your internal APIs)
- Large % of its runs are scheduled rather than explicitly invoked by someone (ambient agents!!)
It became clear we needed one shared place for these primitives, so we built Managed Deep Agents. In my humble opinion it is the easiest way to build a company harness today
Our docs are very agent-friendly, point them here to get started: langch.in/mda -
LangChain
@LangChain
More Managed Deep Agents resources:
🎓 LangChain Academy Module: academy.langchain.com/courses/foundation-introduction-to-…
💻 Managed Deep Agents 101: events.langchain.com/on-demand/f61a1237-3a14-4ff9-9c95-63…
⏯️ YouTube Tutorial playlist: www.youtube.com/watch?v=xdrB53bgpp0&list=PLKYx7JmJ_IcA
LangChain
@LangChain
What is a Managed Deep Agent?
How does it differ from a Deep Agent?
How do you get started?
@jess_weng18 from the LangChain Academy team answers all of these questions in 2 minutesThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Harrison Chase
@hwchase17
deep agents is the harness - the agent loop
managed deep agents adds what you need once lots of people call the same agent: each user's own credentials and their own memory
www.langchain.com/blog/langsmith-managed-deep-agents-what…
LangChain
@LangChain
What is a Managed Deep Agent?
How does it differ from a Deep Agent?
How do you get started?
@jess_weng18 from the LangChain Academy team answers all of these questions in 2 minutesThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
What is a Managed Deep Agent?
How does it differ from a Deep Agent?
How do you get started?
@jess_weng18 from the LangChain Academy team answers all of these questions in 2 minutesThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Human In The Loop: Now in beta in Managed Deep Agents
nathan drezner
@ndrezn
approve / reject isn't a question
sometimes your agent needs to actually ask you something. so let it: drop a few buttons in the slack thread ⏸️ keep going when you click ▶️
public beta in @LangChain Managed Deep Agents -
LangChain
@LangChain
📍 Toronto
📅 Nov 3
We’re teaming up with @TryArcade for a half-day, hands-on workshop.
We'll build a loan-processing agent and take it from first run to governed, observable deployment.
The patterns carry over to any agent that touches customer data, moves money or changes records.
Interested in attending? RSVP below. www.arcade.dev/events/toronto-roadshow-workshop-november-… -
LangChain
@LangChain
As more teams build agents, enterprises need shared ways to develop, evaluate, operate, and govern them.
In this upcoming webinar, we’ll explore the organizational and technical foundations needed to move from isolated agent projects toward a repeatable enterprise approach. -
LangChain reposted
Sydney Runkle
@sydneyrunkle
joined @TalkPython w/ @mkennedy, @reuvenmlerner, @calvinhp, and pat decker to talk about "the state of python in 2026"
so much of it came back to development w/ agents! harnesses, observability, and why ~90% of my coding happens in slack now
talkpython.fm/episodes/show/566/state-of-the-art-in-pytho… -
LangChain
@LangChain
ICYMI: @hwchase17 joined @twentyminutevc with @HarryStebbings to discuss how better models change what you build, where the value in the agent stack ends up, and more.
Watch on YouTube: www.youtube.com/watch?v=1jDCPEHwQfI
Harry Stebbings
@HarryStebbings
What percent of developer salary do you spend on coding tools and how has it changed?
“We spend probably roughly 30% of our engineering salary on coding tools. It has actually gone down the last two months, while the number of tokens that we have used has gone up.
First is visibility, second is basic cost controls, and third is we control the harness.
In that harness, we have done a bunch of work to control costs, things like prompt caching and routing.” @hwchase17
Love to hear your thoughts @bcherny @amasad @HamelHusain @simonwThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
💡 Excellent recap of @hwchase17 + @HarryStebbings' conversation on @twentyminutevc ⤵️
Sydney Runkle
@sydneyrunkle
"it's underrated how hard it is to build agents!"
great convo w/ @hwchase17 and @HarryStebbings on @twentyminutevc
an agent = model + context + harness, and you can improve an agent by tweaking any of the three. a few takeaways:
1. the highest roi thing you can do is start by optimizing context (then the harness, then the weights); this is where domain knowledge lives and that's incredibly valuable
2. the story of continual learning is: run an agent, collect traces, aggregate learnings, fix something, measure the fix, loop! the fix can land in any of the three layers. those fixes all look different but the cycle is the same.
3. memory is still massively unsolved, but also massively important. the hardest part is distilling learnings from traces which is incredibly domain specific.
4. when you build agents, you're probably worried about a few things: accuracy, cost, latency, and security. we've seen these themes confirmed year after year in our dev surveys
5. when we think about reducing cost, the pipeline looks like adding visibility into spend, putting controls in place, and only then pushing on agent optimizations. this is how we reduced our coding agent spend @LangChain
6. model routing, which is great for cost controls, is most effective in the harness (where the context needed for a routing decision lives)
7. open models are used for more narrow tasks right now (small parts of the agent loop OR for very specific domains w/ post-training), but as they improve they'll power more core agent loops
8. we're looking at a future with both horizontal agents and vertical agents. horizontal agents will use the vertical agents, and infra needs to change to support this load (agents can do background work at a rate that humans cannot).
www.youtube.com/watch?v=1jDCPEHwQfI -
LangChain reposted
Sydney Runkle
@sydneyrunkle
"it's underrated how hard it is to build agents!"
great convo w/ @hwchase17 and @HarryStebbings on @twentyminutevc
an agent = model + context + harness, and you can improve an agent by tweaking any of the three. a few takeaways:
1. the highest roi thing you can do is start by optimizing context (then the harness, then the weights); this is where domain knowledge lives and that's incredibly valuable
2. the story of continual learning is: run an agent, collect traces, aggregate learnings, fix something, measure the fix, loop! the fix can land in any of the three layers. those fixes all look different but the cycle is the same.
3. memory is still massively unsolved, but also massively important. the hardest part is distilling learnings from traces which is incredibly domain specific.
4. when you build agents, you're probably worried about a few things: accuracy, cost, latency, and security. we've seen these themes confirmed year after year in our dev surveys
5. when we think about reducing cost, the pipeline looks like adding visibility into spend, putting controls in place, and only then pushing on agent optimizations. this is how we reduced our coding agent spend @LangChain
6. model routing, which is great for cost controls, is most effective in the harness (where the context needed for a routing decision lives)
7. open models are used for more narrow tasks right now (small parts of the agent loop OR for very specific domains w/ post-training), but as they improve they'll power more core agent loops
8. we're looking at a future with both horizontal agents and vertical agents. horizontal agents will use the vertical agents, and infra needs to change to support this load (agents can do background work at a rate that humans cannot).
www.youtube.com/watch?v=1jDCPEHwQfI -
LangChain
@LangChain
Nu by @getnuman is an AI health coach in regulated healthcare, built on LangGraph + LangSmith.
Here's how their team ships fast without crossing the medical device line 🧵
www.langchain.com/blog/how-numan-built-an-eval-first-ai-h…Post not available here, view on X
LangChain
@LangChain
Agent evals run through the same graph.
Every eval, from single-reply fixtures to simulated multi-turn conversations with LLM patient personas, runs against the real agent, so what they score is exactly what a patient would meet.
LangChain
@LangChain
Everything is measured in LangSmith.
1️⃣ All conversations land as a trace, are annotated by reviewers, and feed the loop.
2️⃣ Datasets and experiments live in the same place, allowing them to run themed dataset model configs and compare results.
3️⃣ Triage and improvement workflows read their inputs and write scores back through LangSmith. -
LangChain
@LangChain
🏢 Come to our HQ in SF
🤝 Meet our team
🚀 Kickstart your sales career at LangChain
Don’t miss our SDR recruiting social on October 27th! Save your spot today.
🎟️ RSVP: luma.com/ubhnxwnu -
LangChain
@LangChain
Join our EMEA Deep Life Sci Workshop: Agent Engineering for Life Sciences for a hands-on look at how to build, trace, evaluate, and improve research agents for complex life sciences workflows.
🌍 Register for the EMEA session: events.langchain.com/register/emea-deep-life-sci-workshop/ -
LangChain
@LangChain
✅ Everything we shipped at Interrupt NYC and London
✅ Open source updates
✅ Upcoming events & livestreams
Catch up on everything LangChain in our latest newsletter.
www.langchain.com/blog/october-2026-langchain-newsletter -
LangChain reposted
Sydney Runkle
@sydneyrunkle
harness² from @GoogleResearch offers techniques for RSI on the harness, with a fixed model
they call it "harnessing", but you might already be familiar w/ this process! it's also called "hill climbing" or "continual learning"
more in 🧵 ! -
LangChain
@LangChain
At our recent Learning Loop event, @oneill_c from @baseten explained why continual learning keeps breaking, and what to try.
Watch the full session ⤵️This media couldn't be displayed. See it on the original post. -
LangChain reposted
Sydney Runkle
@sydneyrunkle
sourcing ideas for the next deepagents feature on my morning commute
thanks for the inspo @browserbase -
LangChain reposted
nathan drezner
@ndrezn
it's not easy to properly secure production agents!
we've spent a lot of time building great primitives for Managed Deep Agents to make agents you ship secure by default 🔐
langch.in/mda
Caspar
@caspar_br
A walkthrough of how Managed Deep Agents uses sandbox proxies to keep your credentials safe. langch.in/mda
-
LangChain
@LangChain
Agents with sandboxes run packages you didn't write. If one is malicious, any credential in the sandbox is exposed.
With Managed Deep Agents, the proxy adds your token outside the sandbox, only on calls to services you've allowed. The agent never sees it.
Caspar
@caspar_br
A walkthrough of how Managed Deep Agents uses sandbox proxies to keep your credentials safe. langch.in/mda
-
LangChain
@LangChain
🧵 Behind the scenes of AI Search in LangSmith:
TL;DR
1️⃣ SmithDB turns traces into readable trajectories. Structured filters narrow the candidates when you search
2️⃣ A model checks each remaining trajectory against your description and streams back matchesPost not available here, view on X
LangChain
@LangChain
When you search, structured filters like a time range narrow the candidates first.
Then a small, fast model checks each remaining trajectory against your description, with many checks running in parallel.
Each one usually takes under half second.
Matches stream back as they’re found, so you don’t wait for the full scan.
LangChain
@LangChain
AI Search reads each trajectory within a fixed token budget.
Most questions about behavior only need a small slice, and we checked that on 150 conversations from Open SWE, the coding agent our team uses every day.
✅ The median ran 84K tokens
✅ About 1.5K of that was the user’s messages and the agent’s final reply
✅ Roughly 3/4 was tool output: file reads, shell commands, test logs
www.langchain.com/blog/agent-traces-ai-search-langsmith -
LangChain
@LangChain
What does it take to move healthcare and life sciences agents from experimentation into trusted production systems?
Explore the architectures, evaluation practices, and production patterns teams are using as agents take on more complex work:
➡️ info.langchain.com/guide/scaling-agents-in-healthcare-and… -
LangChain reposted
Sydney Runkle
@sydneyrunkle
we just revamped human in the loop middleware in langchain!
interrupts on tool calls now ship w/ typed schemas, so the human's response is validated before it goes back to the agent
you can also use the schema to render nicer approval forms in your ui
docs.langchain.com/oss/python/langchain/human-in-the-loop -
LangChain
@LangChain
LangSmith for Startups spotlight: @tryopine
The customer intelligence layer for revenue teams
Opine traces every production LLM run through LangSmith, and gates every prompt and model change behind LangSmith evals.
See what Opine looks like on your hardest open deal at tryopine.com.
And if you'd rather help build the context customer intelligence layer than buy it, they’re hiring across GTM and engineering! -
LangChain
@LangChain
Our education team answered the most common questions we've received over the years about agent evals.
Check it out ⤵️
Divya Santhanam
@divyasan0
Agent evals come with a lot of questions, and our team hears the same ones over and over. So I wrote down the most common ones and answered them here ⬇️
www.langchain.com/blog/common-questions-about-agent-evals -
LangChain reposted
Caspar
@caspar_br
Managed Deep Agents can read the files people share and send files back where the conversation already is. It feels like magic
Shoutout to @SlackHQ, their files API made this a joy to build
langch.in/mda -
LangChain
@LangChain
"I can finally find the data that I want without having to fight my tooling."
LangSmith helps @TrustVanta find traces fast, align evaluators to labeled examples, and improve AI quality end to end.
⏯️ Zheng Tao on why "LangSmith gets it."This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Nemotron 3.5 Super VL from @NVIDIAAI brings vision, reasoning & tool use to Deep Agents via ChatNVIDIA.
Open weights. 120B parameters, 12B active.
Nemotron 3.5 Super VL averaged an 87% pass rate on our Deep Agents behavioral suite.
Speed stood out, too. On our Harbor-lite benchmark: 55% less total task time than DeepSeek V4.1 Flash, with similar success.
docs.langchain.com/oss/python/integrations/chat/nvidia_ai… -
LangChain
@LangChain
How do agents remember?
The longer an agent runs, the more it takes in (instructions, tool results, files, and drafts), which results in decreased accuracy and reasoning.
Read how LangChain's Deep Agents harness addresses this by moving context out of the active window without deleting it ⤵️
Jess Weng
@jess_weng18
New blog dropped 👀
www.langchain.com/blog/context-engineering-deep-agents-re… -
LangChain
@LangChain
Claude Platform traces can now be sent to LangSmith through OpenTelemetry.
Export from Claude Managed Agents, the Messages API, or the Claude SDKs with native OTel support. Traces show up in LangSmith so you can inspect every step and improve agent performance.
It’s available now in public beta.
Docs: docs.langchain.com/langsmith/trace-anthropic
ClaudeDevs
@ClaudeDevs
Now in public beta, Claude Platform can send OpenTelemetry traces from the Messages API, Claude Managed Agents, and the Claude SDKs to your observability stack.
This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
How @OrangeBusiness evaluate agents in their enterprise AI platform:
✅ RAG quality: golden datasets, retrieval recall, answer faithfulness
✅ Agent behavior: tool selection accuracy, prompt injection robustness
✅ Production traces: latency, cost, failed runs, human feedback
All in LangSmith, with regression evals before changes reach production, and LangChain for model neutrality & portability
www.langchain.com/blog/how-orange-business-built-live-int…? -
LangChain
@LangChain
ICYMI: AI Search in LangSmith!
LangChain
@LangChain
🔎 Announced at Interrupt: AI Search in LangSmith
You used to need an evaluator or read hours of conversations to find traces that answer questions such as:
💬 Does the user seemed frustrated?"
💬 Where did the agent fail to complete the task?
💬 Where is the user trying to cancel their subscription?
Now, just describe it. LangSmith returns the traces where it happened. www.langchain.com/blog/agent-traces-ai-search-langsmith -
LangChain
@LangChain
And that's a wrap on Interrupt London!
-
LangChain reposted
Sarah B.
@TechWithSarah
being able to search traces in plain language could save teams so much time. especially when you're trying to figure out where an agent went wrong.
LangChain
@LangChain
🔎 Announced at Interrupt: AI Search in LangSmith
You used to need an evaluator or read hours of conversations to find traces that answer questions such as:
💬 Does the user seemed frustrated?"
💬 Where did the agent fail to complete the task?
💬 Where is the user trying to cancel their subscription?
Now, just describe it. LangSmith returns the traces where it happened. www.langchain.com/blog/agent-traces-ai-search-langsmith -
LangChain
@LangChain
The sessions may have ended, but the attendee reception sponsored by @focused_dot_io has just begun. 🥂
-
LangChain reposted
Arjun Natarajan
@arjunnatarajan_
one of the common questions I get is how to find traces based on behaviors of the agent, qualities of the user, or patterns in the execution. Until today that was all either not possible or incredibly convoluted.
I'm super super stoked with our new AI Search feature, try it out!
LangChain
@LangChain
🔎 Announced at Interrupt: AI Search in LangSmith
You used to need an evaluator or read hours of conversations to find traces that answer questions such as:
💬 Does the user seemed frustrated?"
💬 Where did the agent fail to complete the task?
💬 Where is the user trying to cancel their subscription?
Now, just describe it. LangSmith returns the traces where it happened. www.langchain.com/blog/agent-traces-ai-search-langsmith -
LangChain
@LangChain
Final sessions of Interrupt:
🎙️@hwchase17 Fireside Chat with @DunBradstreet SVP, AI Solutions & Data Science Ilya Meyzin
🎙️ Crossing the Production Bottleneck of Agentic Projects with Igor Nikolaienko, Gen AI Architect at @DeutschePostDHL -
LangChain reposted
Simon Budziak
@simon_budziak
@LangChain Interrupt London 2026
Great presentations, food, drinks and people building with AI Agents.
What else do you need? 🤔 -
LangChain
@LangChain
🔎 Announced at Interrupt: AI Search in LangSmith
You used to need an evaluator or read hours of conversations to find traces that answer questions such as:
💬 Does the user seemed frustrated?"
💬 Where did the agent fail to complete the task?
💬 Where is the user trying to cancel their subscription?
Now, just describe it. LangSmith returns the traces where it happened. www.langchain.com/blog/agent-traces-ai-search-langsmith -
LangChain
@LangChain
Train your agent, win a train! 🚂
Stop by the LangChain Agent Arena sponsored by @nvidia to win a LangSmith Engine. -
LangChain
@LangChain
Next up
🎙️@dorco8 at @mondaydotcom on building online & offline LLM Eval Pipelines with LangSmith Traces
🎙️@hwchase17 on building agents that improve over timeThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
🕹️At Interrupt? Swing by the LangChain Agent Arena sponsored by @nvidia
1️⃣ Train an agent by playing Tetris
2️⃣ Choose a model w/ LangSmith LLM Gateway
3️⃣ Watch it compete
🏆 See the results on the leaderboardThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
PSA: We have Interrupt swag 👀
-
LangChain
@LangChain
After lunch:
🎙️How @OVOenergy builds secure customer support agents
🎙️@BenTannyhill on how we built LangSmith Engine
🎙️How @jimdo built an AI companion for small biz with LangSmithThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Sydney Runkle
@sydneyrunkle
we're excited to partner with @scaledown_team to offer support for their new constrained decision models (CDMs) in @LangChain!
these cheap+fast models are tailored to specific critical tasks in the agent loop, like summarization, extraction, and classification.
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
Excited to partner with @scaledown_team 🤝
Their constrained decision models now plug into your agents via:
✅ langchain-scaledown, an OSS package with prebuilt scaledown middlewares
✅ Support in the LangSmith LLM Gateway, with tracing, spend limits, and rate limits
Neal Patel
@neal_k_patel
You already gave the LLM the answer. So why is it writing it out from scratch?
Introducing a new AI model architecture: Constrained Decoding Models (CDMs)
LLMs were built to write. Most production workloads just need an answer: fast, consistent, and affordable.
CDMs are purpose-built for 1 task each. They read your input once and skip the token-by-token generative loop that makes LLMs slow and expensive.
Not just a decision model, but a family of 3 CDMs live today:
• Classify: define your labels and get back a calibrated probability + reason for each. Used for routing, tagging, and intent detection.
• Extract: describe the entities in natural language and get back well-formed JSON with confidence scores + citations. Used for web scraping, document processing, and PII redaction.
• Summarize: pass in long text, images, or PDFs and get back a consistent summary under 1,000 tokens. Used for context compaction, metadata generation, and transcript digests.
Built for production:
• 1M tokens of multimodal (text + image) input, 31x larger than Jev
• 77.2% on DecisionBench vs Jev's 72%
• 200x faster, 467x cheaper than frontier LLM APIs
• $0.05/M input tokens. Output is free (yes, even for summarize)
Live on your favorite platforms today:
@OpenRouter @vercel_dev @LiteLLM @LangChain @TrustedRouter @itsArthurAI @useRouterPlus
Get 50M input tokens for free (for the first 50 hours) at link below 👀 -
LangChain
@LangChain
New in LangChain: Support for @scaledown_team's constrained decision models, which are optimized for quick, cost-effective decisions:
✅ Compatible with LangChain + Deep Agents
✅ Trace every call with LangSmith LLM Gateway
Start summarizing, extracting, and classifying today.
Sydney Runkle
@sydneyrunkle
we're excited to partner with @scaledown_team to offer support for their new constrained decision models (CDMs) in @LangChain!
these cheap+fast models are tailored to specific critical tasks in the agent loop, like summarization, extraction, and classification.
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Marlene Mhangami
@marlene_zw
Made it to another year of @LangChain Interrupt but this time in London 🇬🇧❤️
Just watched @hwchase17’s keynote on owning your intelligence. Outside of GitHub+MSFT ofc, LangChain are still the goats of open source and AI agents, and I’m happy they’ve grown so much!!
Marlene Mhangami
@marlene_zw
Made it for Day 1 of @langchain Interrupt 🥳 So excited to be here!!!
Today is mainly about learning how to build ambient agents 👩🏾💻 -
LangChain
@LangChain
Up next at Interrupt:
🎙️ @GregorBeuster on building visual agents at DB Engineer & Consulting with Deep Agents
🎙️@VictorMoreira16 on what it takes to deploy mission-critical agents to production -
LangChain
@LangChain
LangChain Academy is almost in session!
This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
In between sessions?
Don’t forget to say hello to our sponsors during the show!
@FactoryAI
@p0
@CopilotKit
@tavilyai
@focused_dot_ioThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Simon Budziak
@simon_budziak
@LangChain Interrupt London 2026
Proud to be here as an Ambassador, surrounded by people building the future of AI agents.
At Soba Labs, we're deep in agents and automation.
If you are around let’s grab a coffee and talk ☕
#interrupt26 #langchain #langsmith #aiagents -
LangChain reposted
Julia Schottenstein
@j_schottenstein
London!! @LangChain Interrupt is GLOBAL!
-
LangChain
@LangChain
🔴 @hwchase17’s opening keynote starts NOW!
Join the livestream: interrupt.langchain.com/london/ldn-livestreamThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Not at Interrupt?
We’re livestreaming @hwchase17’s keynote at 10:30am BST.
🔴 interrupt.langchain.com/london/ldn-livestreamThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Starting at 10:30!
First sessions of Interrupt:
🎙️@hwchase17 opening keynote
🎙️Fireside Chat w/ Harrison Chase + @orange Group Chief AI Officer Usman JavaidThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Fuel up for Interrupt!
-
LangChain
@LangChain
The doors are open!
🪪 Grab your badge
☕️ Have a bite
🎉 Get ready for an amazing day
Sessions start at 10:30!This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
POV: You’re arriving at Interrupt London
This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
LangChain has arrived in London!
-
LangChain reposted
Sydney Runkle
@sydneyrunkle
deepagents just hit 30k stars on github! deepagents is great for building agents that do complex work.
they manage context well and use a durable runtime.
though inspired by coding agent design, deepagents are for any domain!
github.com/langchain-ai/deepagents -
LangChain
@LangChain
The 3 questions every agent builder should be able to answer:
✅ Where the agent fails
✅ How to reproduce it
✅ How to make it stop
Watch @JakeBroekhuizen’s session: youtu.be/oVKGKdIAgtgThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
New in Managed Deep Agents 0.9: Reactions API for @SlackHQ Channels.
When it comes to building effective user agent experiences, loading states are underrated.
Now, you can build loading states and receipts using either heuristics, decision models, or custom logic.
nathan drezner
@ndrezn
We've been thinking a lot about great agent UX with Managed Deep Agents - here's some of what we've been building in this area and a bit about our product philosophy x.com/i/article/2108561070809821184
-
LangChain
@LangChain
Experience the best of Interrupt.
Access our archive and watch every session on-demand: interrupt.langchain.com/recordings
LangChain
@LangChain
We simplified agent authentication, memory, and channels, and brought in web search as a pre-built tool for Managed Deep Agents.
Watch @VictorMoreira16’s full Interrupt NYC session for an inside look: youtu.be/XcsspzECTKAThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
We simplified agent authentication, memory, and channels, and brought in web search as a pre-built tool for Managed Deep Agents.
Watch @VictorMoreira16’s full Interrupt NYC session for an inside look: youtu.be/XcsspzECTKAThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
LangSmith Signals from the last month:
✅ Claude Sonnet 5: #9 to #2 in model adoption. 51% more orgs using it
✅ gpt 5.6 luna: #3 to #1 in call footprints. 65% more calls
✅ Two open-weights models are now in the adoption top 10, but not for call volume
✅ Smaller, faster models dominate call footprints -
LangChain
@LangChain
✅ 60k queries handled
✅ 500+ customer accounts
✅ 85%+ of sessions resolved without a support ticket
✅ 250+ cases auto-detected + escalated to the right team
@SnykSec Assist. Built on LangChain + LangGraph. Observability in LangSmith.
Full story: www.langchain.com/blog/how-snyk-turned-an-internal-suppor… -
LangChain
@LangChain
.@Factory_AI is sponsoring Interrupt London!
Stop by their booth during the event and go hands-on with the autonomy stack for enterprise teams.
LangChain
@LangChain
Learn more about Interrupt, The Agent Conference by LangChain and RSVP: interrupt.langchain.com/london
-
LangChain
@LangChain
For most global teams, flexibility matters as agent workloads grow, architectures evolve, and infrastructure requirements become more complex.
Join @AWScloud, @NVIDIA, and LangChain to learn how teams can approach the infrastructure layer behind production agent systems.
Save your spot: aws-experience.com/apj/smb/e/0700a/architecting-flexible-… -
LangChain reposted
Harrison Chase
@hwchase17
Sam on where decision models fit in a harness and how to use Jev with LangChain
www.langchain.com/blog/building-a-harness-with-jev
Jamin Ball
@jaminball
Next up on First Pass: a conversation with
@samecrowder from @LangChain
on decision models.
Jev from @typesafeai took the world by storm. Then OpenAI and Databricks released producs (Decisions API and ai_decide function) So what are Decision Models? We dig in!This media couldn't be displayed. See it on the original post. -
LangChain reposted
Sydney Runkle
@sydneyrunkle
the @OpenAI decisions API is now in @LangChain!
we shipped:
1. OpenAIDecisions: ask yes/no, choice, and score questions
2. model routing middleware
3. auto mode middleware
docs: docs.langchain.com/oss/python/integrations/decision_model… -
LangChain reposted
Jamin Ball
@jaminball
Next up on First Pass: a conversation with
@samecrowder from @LangChain
on decision models.
Jev from @typesafeai took the world by storm. Then OpenAI and Databricks released producs (Decisions API and ai_decide function) So what are Decision Models? We dig in!This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Here’s an app we built called Restock, an office supply agent that works in @SlackHQ.
A quick demo ⏯️This media couldn't be displayed. See it on the original post.
LangChain
@LangChain
You can now build an agent that:
🔎 Finds real products
🛒 Prepares purchases
💳 Pays
…with a person approving every order 👨💻
Ask in @slackhq, review the order, approve in @Stripe's Link agent wallet.
Built on MPP + Managed Deep Agents.
www.langchain.com/blog/agents-that-can-pay-with-stripe-link -
LangChain reposted
Jeff Weinstein
@jeff_weinstein
LangChain developers can now easily build agents that can pay for things with @link (which works across any checkout online or programmatic payments over @mpp).
See their in-depth guide and demo! =>
LangChain
@LangChain
You can now build an agent that:
🔎 Finds real products
🛒 Prepares purchases
💳 Pays
…with a person approving every order 👨💻
Ask in @slackhq, review the order, approve in @Stripe's Link agent wallet.
Built on MPP + Managed Deep Agents.
www.langchain.com/blog/agents-that-can-pay-with-stripe-link -
LangChain
@LangChain
RT @ndrezn: we built Restock, an agent to automate ordering office supplies -pens, groceries, and more. it's open source, so you can fork i…
LangChain
@LangChain
You can now build an agent that:
🔎 Finds real products
🛒 Prepares purchases
💳 Pays
…with a person approving every order 👨💻
Ask in @slackhq, review the order, approve in @Stripe's Link agent wallet.
Built on MPP + Managed Deep Agents.
www.langchain.com/blog/agents-that-can-pay-with-stripe-link -
LangChain reposted
Sydney Runkle
@sydneyrunkle
now @link from @stripe makes it really easy to build agents that can make payments, with human approval on every purchase
agents like muse and instinct popularized payments for personal agents, MDA makes this easy to do for enterprise
LangChain
@LangChain
You can now build an agent that:
🔎 Finds real products
🛒 Prepares purchases
💳 Pays
…with a person approving every order 👨💻
Ask in @slackhq, review the order, approve in @Stripe's Link agent wallet.
Built on MPP + Managed Deep Agents.
www.langchain.com/blog/agents-that-can-pay-with-stripe-link -
LangChain
@LangChain
You can now build an agent that:
🔎 Finds real products
🛒 Prepares purchases
💳 Pays
…with a person approving every order 👨💻
Ask in @slackhq, review the order, approve in @Stripe's Link agent wallet.
Built on MPP + Managed Deep Agents.
www.langchain.com/blog/agents-that-can-pay-with-stripe-link -
LangChain
@LangChain
At #sftechweek? Join LangChain, @tavilyai, and @nebius for an afternoon bringing together enterprise leaders, founders, and engineers building with the agentic layer.
Info + RSVP: luma.com/tavily-gbt7 -
LangChain
@LangChain
At Interrupt, The Agent Conference by LangChain, @OVOEnergy will be sharing how they built a flywheel to improve customer support agents.
Catch OVO's talk + more: interrupt.langchain.com/london -
LangChain reposted
Sydney Runkle
@sydneyrunkle
i'm most excited about the idea of "binding tools to skills"
this is basically progressive disclosure^2 -- when a skill is activated w/ relevant instructions, it comes w/ corresponding tools
great feature if you want to assemble context just in time for a given task
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
LangChain has ranked #12 on @Paraform’s Talent Density Index.
As a growing team of builders making an outsized impact in our industry, we are proud to receive this recognition.
Want to join us? We’re hiring! www.langchain.com/careers -
LangChain reposted
Harrison Chase
@hwchase17
interrupt london is next tuesday
really looking forward to hearing from Orange and OVO on what's actually working with agents in production. we're also running hands-on workshops on Deep Agents and Engine
LangChain
@LangChain
Only one week to go until Interrupt London!
interrupt.langchain.com/london -
LangChain reposted
Harrison Chase
@hwchase17
One cool thing deepagents now supports: dynamically loading tools when is a skill is loaded
Some skills require specific tools that you may not always want to have available. This now allows you to pair them
With openai/anthropic models, it’s possible to do this in a way where it doesn’t break the prompt cache!
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
Teams are narrowing in on skills as the standard for providing domain knowledge to agents.
Here's how we've revamped skills in Deep Agents to meet growing demand!
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Caspar
@caspar_br
managed deep agents 0.9 is here! things we shipped:
⏰ agents can schedule tasks OBO themselves or the user
🏭 dynamically configure your mda per run (models, tools, skills, etc.)
👀 Slack reactions show when your agent has seen a message
📁 improved multimodal capabilities
🔌 MCP 2.0 support via langchain.mcp
released for js and py langch.in/mda -
LangChain
@LangChain
Managed Deep Agents v0.9 is here!
✅ Schedules SDK: Agents create their own reminders, follow-ups, and recurring tasks mid-conversation
✅ Per-run config: Pick the model, skills, MCP servers, and sandbox for each run, so one deployment serves every team or repo
✅ Slack Reactions: Agents react to messages as soon as they start a run www.langchain.com/blog/managed-deep-agents-schedules-per-… -
LangChain
@LangChain
What is an agent skill? (Or how to protect your agent’s context window)
A 2 minute explainer from LangChain Academy.
🎓 Take our course on Deep Agents: academy.langchain.com/courses/foundation-introduction-to-…This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
✅ Red team your agents
✅ Detect more issue types
✅ Automatically test proposed fixes
@bentannyhill on everything new in LangSmith Engine v2 and how we approach agent improvement.
Watch the full Interrupt NYC Session: youtu.be/S3l7U7gOsyYThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Don't Miss @hwchase17's Interrupt London Keynote
Airing October 13th
🔴 RSVP: interrupt.langchain.com/london/ldn-livestream -
LangChain
@LangChain
.@tavilyai is sponsoring Interrupt London!
Stop by their booth during the event and see how you can connect your AI agents to the web.
LangChain
@LangChain
Learn more about Interrupt, The Agent Conference by LangChain and RSVP: interrupt.langchain.com/london
-
LangChain
@LangChain
Deep Agents explained in 90 seconds.
This media couldn't be displayed. See it on the original post.
LangChain
@LangChain
Learn to build long-running agents for complex workflows with Deep Agents:
🎓 Take our LangChain Academy Course today.
academy.langchain.com/courses/foundation-introduction-to-… -
LangChain reposted
Caspar
@caspar_br
let me introduce you to the langchain stack
1. langgraph gives you this level of control
2. traces give you the complete run lineage
3. version your skills and instructions in context hub
4 & 5. deepagents has durable execution built in and manages the context window for you
Josh Rosen
@JoshARosen
We should be stealing a lot more from data engineering when we build AI systems.
DAGs - does this really need to be one giant agent loop?
Lineage - which sources, artifacts, prompts, and model versions produced this?
Change detection - if a source changes, which generated artifacts are now outdated?
Incremental processing - what can we keep instead of rerunning the whole agent?
ETL - what context do we pull in, what does the model do with it, and where do we save the result?
Data engineers have been solving these problems for decades. -
LangChain
@LangChain
New vid from @sydneyrunkle on how to build a model router into the harness:
1️⃣ Understand the tasks
2️⃣ Understand the models
3️⃣ Build the router in the harness
4️⃣ Track task outcomes
Sydney Runkle
@sydneyrunkle
not every task needs a frontier model!
we built a model router into our coding agent's harness and cut costs by 64% w/o reducing quality.
this vid walks through how it works, and how you can build one into your own harness!This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
It's super effective!
Caspar
@caspar_br
time to level up langch.in/mda
This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
The agentic operating model helps enterprise teams understand questions like:
✅ Who owns agent development and operations?
✅ What should be standardized across teams?
✅ Where should governance and controls live?
✅ How do teams share infrastructure without slowing down development?
✅ How do you continuously evaluate and improve production agents?
Join our upcoming webinar, The Agentic Operating Model, to explore how people, process, and technology need to evolve together as organizations scale production AI.
Register: events.langchain.com/webinar/emea-the-agentic-operating-m… -
LangChain reposted
nathan drezner
@ndrezn
free web search! for your agents! backed by Parallel with ZDR. Take it for a spin today.
LangChain
@LangChain
Add web search to your Managed Deep Agents for free.
🌐 Powered by @p0
💻 No separate vendor accounts, managing extra PI keys, or wiring search tools necessary
A quick demo from @ndreznThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Add web search to your Managed Deep Agents for free.
🌐 Powered by @p0
💻 No separate vendor accounts, managing extra PI keys, or wiring search tools necessary
A quick demo from @ndreznThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
.@Rippling AI: Built On Deep Agents + LangSmith
The architecture:
✅ A supervisor agent coordinating 5 to 7 specialized subagents
✅ Tracing, evaluations, and production monitoring handled by LangSmith
www.langchain.com/blog/how-rippling-went-ai-native-across… -
LangChain
@LangChain
Only one week to go until Interrupt London!
interrupt.langchain.com/london -
LangChain reposted
Sydney Runkle
@sydneyrunkle
deep agents just shipped a round of multimodal upgrades!
1. broader file support on OpenAI Responses: docx, pptx, xlsx, csv, json, markdown, not just pdfs
2. opt-in binary offloading to keep big files out of your checkpoints
3. a batch of reliability fixes
more in thread 🧵 ! -
LangChain
@LangChain
Adding a new feature to @theziphq's agent used to take weeks. Their team would have to hand-build the code for LLM calls, storage & context.
They rebuilt their chatbot on LangGraph + LangSmith, and every trace was there without them writing a single line of code.
Full story ⤵️This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
🗽 Every session from Interrupt NYC, now on demand.
⏯️ Access the archive interrupt.langchain.com/recordings
LangChain
@LangChain
At Interrupt NYC, @hwchase17 kicked off our agent conference with a session on owning your intelligence, why so many teams care about it right now, and our latest launches.
⏯️ Watch the full session: youtu.be/950byF7njfwThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
At Interrupt NYC, @hwchase17 kicked off our agent conference with a session on owning your intelligence, why so many teams care about it right now, and our latest launches.
⏯️ Watch the full session: youtu.be/950byF7njfwThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
In our Deep Life Sci Workshop, we’ll work through a real research-agent workflow and explore how teams can:
✅ Connect agents to scientific data and research sources
✅ Trace agent reasoning and tool usage
✅ Evaluate agent performance
✅ Identify and fix issues using traces
✅ Improve complex research workflows over time
🌎 NAMER: events.langchain.com/webinar/namer-deep-life-sci-workshop/
🌍 EMEA: events.langchain.com/register/emea-deep-life-sci-workshop/ -
LangChain
@LangChain
.@p0 is sponsoring Interrupt London!
Stop by their booth for a live demo of the infrastructure that lets AI search, extract, monitor, and reason over the web.
LangChain
@LangChain
Learn more about Interrupt, The Agent Conference by LangChain and RSVP: interrupt.langchain.com/london
-
LangChain
@LangChain
Attending @WorldSummitAI?
⛴️ Don’t miss our evening tasting tour of Amsterdam by boat and by foot featuring @Elastic and @GoogleCloud on October 7th 🚶
Space is limited, RSVP today!
luma.com/20qfqog2 -
LangChain
@LangChain
Experience the best of Interrupt NYC, now on demand.
interrupt.langchain.com/recordings -
LangChain reposted
Sydney Runkle
@sydneyrunkle
user memory is what helps agents improve over time!
for example, our content studio agent (built on MDA, made the video below, s/o @caspar_br) uses user level memory.
it can:
- save my preferences when i give it feedback (like slowing down text animations)
- remember our past conversations, including previous content artifacts
so not only does the content improve over time but it's also easier/faster to generate!
Caspar
@caspar_br
User memory in Managed Deep Agents 0.8: your agent can remember the people it works with langch.in/mda
This media couldn't be displayed. See it on the original post. -
LangChain reposted
Caspar
@caspar_br
User memory in Managed Deep Agents 0.8: your agent can remember the people it works with langch.in/mda
This media couldn't be displayed. See it on the original post. -
LangChain reposted
Sydney Runkle
@sydneyrunkle
another interesting lesson from this is that you should always be experimenting w/ cheaper models, you'd be surprised at how good open / lower tier models are!
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
💻 Build the interface you want w/ your LangSmith data
✅ Publish it into your workspace
🙅 Skip hosting, auth, + permissions work
Get started with Custom Apps today.
www.langchain.com/blog/langsmith-custom-apps -
LangChain
@LangChain
Scope agent memory to a single user or your entire team.
Managed Deep Agents now supports user-level memory out of the box. Set it up with one file and deploy with a single command.
Caspar
@caspar_br
User memory in Managed Deep Agents 0.8: your agent can remember the people it works with langch.in/mda
This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Validate agent fixes before shipping them with LangSmith Engine v2.
Engine now tests and validates fixes automatically:
✅ Replicates the issue with the same deployment environment and same inputs
✅ Builds a fix, tests it, and iterates on improving it until it has a satisfactory solution
✅ You review the fix and deploy it with a ready-made PR
LangChain
@LangChain
Learn more → www.langchain.com/blog/langsmith-engine-v2-redteam
-
LangChain
@LangChain
Ilya Meyzin, SVP, AI Solutions & Data Science at @DunBradstreet is speaking at Interrupt, The Agent Conference by LangChain.
Catch Ilya’s talk + more: interrupt.langchain.com/london -
LangChain
@LangChain
LangSmith for Startups Spotlight: @corridor
@Corridor uses LangSmith to develop + run novel agentic security evaluations, build + trace inference pipelines, and manage memory graphs that provide specialized context across various product surfaces.
Get a demo: -
LangChain reposted
Hunter Lovell
@huntlovell
some much needed TLC for langchain mcp!
LangChain JS
@LangChain_JS
@𝚕𝚊𝚗𝚐𝚌𝚑𝚊𝚒𝚗/𝚖𝚌𝚙-𝚊𝚍𝚊𝚙𝚝𝚎𝚛𝚜 2.0 is out! 🧵
Using MCP with your TypeScript agents up to MCP servers just got a lot simpler: support for the latest stateless version of MCP, tools that can check in with your users when they need something, much easier auth integrations, and a ton of QoL updates.
Here's what's new 👇 -
LangChain
@LangChain
🧵 @langchain / MCP-adapters 2.0
LangChain JS
@LangChain_JS
@𝚕𝚊𝚗𝚐𝚌𝚑𝚊𝚒𝚗/𝚖𝚌𝚙-𝚊𝚍𝚊𝚙𝚝𝚎𝚛𝚜 2.0 is out! 🧵
Using MCP with your TypeScript agents up to MCP servers just got a lot simpler: support for the latest stateless version of MCP, tools that can check in with your users when they need something, much easier auth integrations, and a ton of QoL updates.
Here's what's new 👇 -
LangChain
@LangChain
LangChain Academy Course Update: Introduction to Deep Agents.
Deep Agents makes it easier to build agents, Managed Deep Agents lets you deploy them with a single CLI command. We introduce Managed Deep Agents and show you how to deploy an agent in Slack.
Check it out today: -
LangChain reposted
Harrison Chase
@hwchase17
model routing - "no one knows what it means, but its provocative, gets the people going"
we try to shed light on how to do it properly
1. understand tasks
2. understand the models
3. build the router **in the harness**
4. track outcomes
lower costs, with no performance hit
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
nathan drezner
@ndrezn
great piece from @sydneyrunkle on moving model routing upstream into the harness -
great way to think about managing routing more efficiently and building your harness to avoid context bloat.
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Sydney Runkle
@sydneyrunkle
we built a model router for our coding agent that cut median cost per task by 64%, with no measurable drop in quality!
most tasks don't need top tier intelligence!
here's how we built it, why we think it belongs in the harness, and how to add one to your own agent :)
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
Adding routing to your agent? Here's where to start:
1️⃣ Understand the tasks
2️⃣ Understand the models
3️⃣ Build the router in the harness
4️⃣ Track task outcomes
An inside look at how to build a model router in the harness, and how we cut our median cost per thread by 64% with no change in quality.
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Viv
@Vtrivedy10
we want every team to own this cycle:
- log every agent action in Trajectories
- use agents to help humans understand that data at scale
- turn that data into Evals & Environments
- do lots of Agent + Human review
- Fine-tune much cheaper and faster open models that rock at those tasks
LangChain
@LangChain
Last week, we launched LangSmith Fine-Tuning and the smithtune CLI.
Turn your agent’s own trajectories into a fine-tuned model without building a training pipeline by hand.
Your traces already show your agent doing the job well. Now, that signal becomes training data.
Quick demo from @jakebroekhuizenThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Ankush Gola
@ankush_gola11
A huge part of post-training is picking and wrangling the right data. `smithtune` handles the entire lifecycle of agent traces -> fine-tuned model
Jake Broekhuizen
@jakebroekhuizen
We spent a lot of time streamlining the process to go from your agent's trajectories to an SFT'd model on your training/inference provider of choice so you don't have to stitch together the steps manually.
The data curation phase is definitely the most important, and we've had great results from working with training copilots in the form of model councils and a coding agent guided by the CLI and its packaged skills. This + the good ol' fashioned 'look at the data' ring true for any well-designed eval & post-training pipeline, and we're really excited to enable builders to use this in the process of owning your own intelligence!
LangChain
@LangChain
Last week, we launched LangSmith Fine-Tuning and the smithtune CLI.
Turn your agent’s own trajectories into a fine-tuned model without building a training pipeline by hand.
Your traces already show your agent doing the job well. Now, that signal becomes training data.
Quick demo from @jakebroekhuizenThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Last week, we launched LangSmith Fine-Tuning and the smithtune CLI.
Turn your agent’s own trajectories into a fine-tuned model without building a training pipeline by hand.
Your traces already show your agent doing the job well. Now, that signal becomes training data.
Quick demo from @jakebroekhuizenThis media couldn't be displayed. See it on the original post.
LangChain
@LangChain
We ran it on two of our own agents.
OpenSWE Review's precision went from 62.9% to 81.5% with 29% fewer tool calls per review.
Engine, fine-tuned on Kimi K3, beat the base model and GPT-5.6 Sol on our issue-detection benchmark.
www.langchain.com/blog/langsmith-fine-tuning -
LangChain
@LangChain
.@AbridgeHQ built clinician feedback into an evaluation system that helped the team measure agent quality, identify improvements, and support faster releases.
It’s one example of how healthcare organizations are moving beyond subjective review toward repeatable systems for evaluating production AI.
See more in our free guide: info.langchain.com/guide/scaling-agents-in-healthcare-and… -
LangChain
@LangChain
.@CopilotKit is sponsoring Interrupt London!
Stop by their booth during the event and go hands-on with the enterprise-ready frontend stack for agents.
LangChain
@LangChain
Learn more about Interrupt, The Agent Conference by LangChain and RSVP: interrupt.langchain.com/london
-
LangChain
@LangChain
Headed to @WorldSummitAI?
👥 Meet the LangChain team at Booth G54
🎙️At 11:50am on Oct 7, catch @hwchase17’s talk on the state of agent intelligence
⛴️ After Day 1, set sail with LangChain + @Elastic + @googlecloud luma.com/20qfqog2
💻 At 11:50am on Oct 8, stop by our Agent Development workshop with LangChain & LangSmith -
LangChain reposted
Sydney Runkle
@sydneyrunkle
it's really hard to "score" a PR but "human touches per PR" is quite easy to track
one complicating factor here though is even if anagent is doing well, on a complex task, more touches isn't necessarily bad
we've been using this metric for evaluating our coding agents!
Zach Lloyd
@zachlloydtweets
Software engineering is shifting to "factory engineering:" the focus isn't just on building the product, but improving throughput.
One measure of factory efficiency is "human touches per pr." Over time you want to drive this down.This media couldn't be displayed. See it on the original post. -
LangChain reposted
Sydney Runkle
@sydneyrunkle
this is cool
* i trigger a coding agent in slack
* i review PRs in github, leaving comments
* my coding agent automatically triggered and addresses my comments
* PRs approved, ofc!
so this is what a software factory feels like :) -
LangChain
@LangChain
.@sydneyrunkle on the latest in LangGraph
Sydney Runkle
@sydneyrunkle
new in langgraph: interrupts with typed response schemas!
when your agent pauses for a human, you define what a valid response looks like:
1. responses are validated before they're passed back to the agent
2. you can render a real response form in your UI, not just a json entry field
human in the loop, but (type) safe
docs: docs.langchain.com/oss/python/langgraph/interrupts -
LangChain reposted
Sydney Runkle
@sydneyrunkle
i only use agents if they're highly convenient / available (mostly in slack these days)
HTTP channels are perfect for enabling this, great explainer from @ndrezn
LangChain
@LangChain
Managed Deep Agents 0.8 ships HTTP channels: agents can take requests from any webhook, not just Slack.
@ndrezn built an expense agent that lives in text messages, backed by @PhotonHQ.
Text a receipt, get it logged automatically.
More on Managed Deep Agents:This media couldn't be displayed. See it on the original post. -
LangChain reposted
matt
@matt_feroz
I filmed an episode of Agent Clinic with @GoogleCloudTech!
We built an automated eval suite for a docs agent I built in @LangChain using @merge_api Gateway as the intelligence provider.
Shoutout to the three OSS projects I ran my agent on:
@t3dotcodes by @theo + @jullerino
@opencode by @thdxr
@pidotdev by @badlogicgamesThis media couldn't be displayed. See it on the original post.
Google Cloud Tech
@GoogleCloudTech
Terminal test runs won't catch multi-turn agent regressions.
In Ep 3 of Agent Clinic, Dani Zamora & @matt_feroz build an automated eval suite for a LangGraph agent in 60 mins.
Here's the 4-step framework to go from vibes to benchmark ANY AI agent → goo.gle/4hAkdcz -
LangChain reposted
Sydney Runkle
@sydneyrunkle
new in langgraph: interrupts with typed response schemas!
when your agent pauses for a human, you define what a valid response looks like:
1. responses are validated before they're passed back to the agent
2. you can render a real response form in your UI, not just a json entry field
human in the loop, but (type) safe
docs: docs.langchain.com/oss/python/langgraph/interrupts -
LangChain
@LangChain
Headed to #sftechweek?
We’re partnering with @modal + @cogent_security to host a live, team-based AI Heist challenge.
You'll work in a fictional, isolated cloud environment, direct an agent through staged security challenges to find the hidden proof token, and compete on a live leaderboard for prizes.
Capacity is limited. RSVP today! partiful.com/e/CQdFO5si7fnkMcfXet7q -
LangChain reposted
nathan drezner
@ndrezn
I had some folks asking for the source code for this demo!
Link in 🧵 if you want to fork/build off of it in any direction. Our documentation for HTTP channels is also fantastic.
LangChain
@LangChain
Managed Deep Agents 0.8 ships HTTP channels: agents can take requests from any webhook, not just Slack.
@ndrezn built an expense agent that lives in text messages, backed by @PhotonHQ.
Text a receipt, get it logged automatically.
More on Managed Deep Agents:This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
The LangSmith Roadshow continues tomorrow in Los Angeles!
✅ Learn about the state of agent development, and what it takes to ship to production in 2026
✅ Go hands-on with LangSmith Engine and Deep Agents during our conceptual deep dives
✅ Hear how @ServiceTitan uses LangSmith in production
🪄 And don’t miss the live magic show at the end!
Spaces are limited! RSVP at events.langchain.com/LangSmithRoadshow/LosAngeles/ -
LangChain
@LangChain
Managed Deep Agents 0.8 ships HTTP channels: agents can take requests from any webhook, not just Slack.
@ndrezn built an expense agent that lives in text messages, backed by @PhotonHQ.
Text a receipt, get it logged automatically.
More on Managed Deep Agents:This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Spot agent issues before they’ve appeared in production with LangSmith Engine v2.
Engine now red teams your agent proactively 👇
LangChain
@LangChain
Engine learns how your agent works by reading traces and repos, then tests for agent-specific weaknesses and flags any it confirms.
You get a list of verified issues – from hallucinations to violations of system prompts – so you can fix them before they impact your users.
LangChain
@LangChain
Learn more → www.langchain.com/blog/langsmith-engine-v2-redteam
-
LangChain
@LangChain
What does it take to move from isolated agent experiments to production AI at enterprise scale?
Our latest guide brings together lessons from @SchneiderElec, V@odafone, and @mondaydotcom across:
- Shared agent platforms and LLMOps
- Production observability and evaluation
- Multi-agent architectures
- Clearer boundaries between agents, tools, and environments
- Security, governance, and data residency
- Scaling adoption across teams and business units
info.langchain.com/guide/scaling-agents-in-europe-and-the… -
LangChain
@LangChain
Igor Nikolaienko, AI Architect at @DeutschePostDHL is taking the stage at Interrupt, the Agent Conference by LangChain.
Catch Igor’s talk + more: interrupt.langchain.com/london -
LangChain
@LangChain
A customer says the agent got it wrong.
LangSmith is how @Clay finds out why.This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Trajectories in LangSmith are a chronological, human-readable view of an agent session.
✅ Every run rolls into a trace, and traces from a session link into a thread: the full execution tree, nested runs and all.
✅ A trajectory strips that out, keeping each message once, in order, so you read the session as the path the agent took.
LangChain
@LangChain
✅ Debug long-running agents faster, skip the reconstruction
✅ Give SMEs a readable view to review agent behavior
✅ Score agent behavior with online evals
✅ Turn production sessions into fine-tuning datasets
Try Trajectories in LangSmith today.
www.langchain.com/blog/langsmith-trajectories-tracing -
LangChain
@LangChain
Packed house at the LangSmith Roadshow Boston today!
📍Find one near you: events.langchain.com/LangSmithRoadshow/ -
LangChain
@LangChain
Meet the new terminal interface in Managed Deep Agents.
Get started: `mda chat`
nathan drezner
@ndrezn
We've just shipped a new terminal interface 💻 for Managed Deep Agents. It rips.
Use it to instantly test drive your agent locally. Studio isn't going anywhere; the TUI is handy for lightweight development.
Take it for a spin with `mda chat`.
Personally I'm a big fan of the theming the team put together 🌈 -
LangChain
@LangChain
LangChain 🤝 @Box
Box
@Box
Agentic orchestration, governed content, structured extraction, auditable logic, and human approval in one flow. Check out how @LangChain and Box make that possible. 👇 x.com/i/article/2103554266300628992
-
LangChain
@LangChain
Join us for the Deep Life Sci Workshop: Agent Engineering for Life Sciences, and learn about the agent engineering patterns behind more accurate, observable, and reliable research workflows.
🌎 NAMER: events.langchain.com/webinar/namer-deep-life-sci-workshop/
🌍 EMEA: events.langchain.com/register/emea-deep-life-sci-workshop/ -
LangChain
@LangChain
.@focused_dot_io is sponsoring Interrupt London!
Stop by their booth during the event and meet the trusted LangChain partners building agents that integrate with your existing systems, and don’t miss the attendee reception at the end of Interrupt sponsored by their team!
LangChain
@LangChain
Learn more about Interrupt, The Agent Conference by LangChain and RSVP: interrupt.langchain.com/london
-
LangChain
@LangChain
This just in: @sydneyrunkle sat down with @jlowin and @zzstoatzz from @PrefectIO to talk about the new MCP spec + FastMCP, Jev, evals, and more!
Give it a listen!
www.youtube.com/watch?v=AuAwd5TF_lg -
LangChain reposted
Sydney Runkle
@sydneyrunkle
"our feedback loop went from hours and hours to... automated, just because of langsmith"
@ktech9999 is one of my favorite builders. he lives at the bleeding edge, and his agents are ridiculously effective
learn about one of his favorite tools for automating agent improvement👇
LangChain
@LangChain
🍽️ @wonder is building an agent that plans, orders and ensures delivery of all 21 of your weekly meals, while handling all customizations and customer service resolutions.
With LangSmith, they swapped log debugging for LLM-as-judge scoring. A PM can file a bug and Claude Code fixes it from the trace.
⏯️ @ktech9999’s storyThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Sydney Runkle
@sydneyrunkle
had a great convo w/ @jlowin and @zzstoatzz on lots of hot topics in the agent space:
* the new MCP spec, and how to use it w/ FastMCP and @LangChain!
* evals (and why they're hard)
* decision models like jev
Prefect
@PrefectIO
fastmcp pod is up! in this episode, @jlowin and @zzstoatzz sat down with @sydneyrunkle from @LangChain to talk mcp, building agents, and how ai is changing open source.
they get into 👇
• why langchain chose fastmcp to connect agents to mcp servers
• keeping open source collaborative when maintainers would rather use their own coding agents to write the code
• how to tell whether an mcp tool actually helps an agent
• why sydney thinks agent development could look mostly like designing evals a year from now
• experiments with jev for tool selection, model routing, and evals
watch the full episode ↓This media couldn't be displayed. See it on the original post. -
LangChain reposted
Kartik Arora 🔋
@ktech9999
If you’re building AI agents, I would highly recommend Langchain and Langsmith.
LangChain
@LangChain
🍽️ @wonder is building an agent that plans, orders and ensures delivery of all 21 of your weekly meals, while handling all customizations and customer service resolutions.
With LangSmith, they swapped log debugging for LLM-as-judge scoring. A PM can file a bug and Claude Code fixes it from the trace.
⏯️ @ktech9999’s storyThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
🍽️ @wonder is building an agent that plans, orders and ensures delivery of all 21 of your weekly meals, while handling all customizations and customer service resolutions.
With LangSmith, they swapped log debugging for LLM-as-judge scoring. A PM can file a bug and Claude Code fixes it from the trace.
⏯️ @ktech9999’s storyThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Proud to be a launch partner on @nvidia's new Open Agent Safety Platform.
Jensen Huang
@JensenHuang
Today, with over 100 industry partners, we introduced the NVIDIA Open Agent Safety Platform, bringing together OpenShell and Sentry.
Artificial intelligence is extraordinary technology that will advance discovery, productivity, security, health, and prosperity for generations to come.
But its full promise can only be realized when people have confidence that AI is being built to be safe and deployed with wisdom and responsibility.
This is bigger than a single product. It's the beginning of an open ecosystem to build the trust layer for safe agent systems.
Together, we are building the foundation of the AI economy.
Trust and innovation are not in conflict. Safety is how trust is earned. We must build not only the most capable AI, but the most trusted AI, so that this extraordinary technology can realize its enormous promise for the world. nvda.ws/4hcoq7m -
LangChain
@LangChain
With LangSmith Engine v2, you can:
1️⃣ Proactively detect issues before they appear in production with Red Teaming
2️⃣ Quickly resolve agent issues with fixes validated in your environment
3️⃣ Detect harder-to-spot issues like inefficient agent work and metric trends -
LangChain
@LangChain
What separates a simple agent from one that can reliably work through complex, long-running tasks?
It comes down to the harness around the model: how the agent plans, manages context, uses tools, and breaks work into smaller steps.
In our upcoming virtual workshop, we’ll walk through these patterns and give you a practical look at how Deep Agents are built.
Choose the session that works best for your region:
🌍 EMEA: events.langchain.com/register/emea-agent-harnesses-and-de…
🌏 APAC + ANZ: events.langchain.com/register/apac-anz-agent-harnesses-an…
LangChain
@LangChain
Choose the session that works best for your region:
🌍 EMEA: events.langchain.com/register/emea-agent-harnesses-and-de…
🌏 APAC + ANZ: events.langchain.com/register/apac-anz-agent-harnesses-an… -
LangChain
@LangChain
Usman Javaid, Group Chief AI Officer at @orange is headlining Interrupt, The Agent Conference by LangChain.
Catch his talk + more on October 13th. interrupt.langchain.com/london -
LangChain reposted
Hunter Lovell
@huntlovell
jev was really exciting to us since it’s a reprisal of how we think a lot of systems with AI should be built!
think about it this way: when starting a project you don’t start with big behemoth k8s clusters (à la frontier models/ harnesses), you start with smaller units (jev) and iterate as needed
we learned A LOT when simple decisions from LLMs were the paradigm (which is what langgraph was built around), to see this type of model stack rank next to models of today (*on subset of tasks) makes us really excited for the future
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Harrison Chase
@hwchase17
jev and langgraph are a great combo!
modeling agents as complex systems, with AI imbued in them, makes a ton of sense
langgraph is the best way to model them as such, and using Jev inside langgraph to turbo charge all the small decisions that need to be made is awesome combo
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
Jev 🤝 LangGraph
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
1️⃣ LangSmith Engine v2
2️⃣ Managed Deep Agents v0.8
3️⃣ LangSmith Trajectories
4️⃣ LangSmith Fine-Tuning
5️⃣ Custom Apps
Everything we announced at Interrupt NYC ⤵️
www.langchain.com/blog/langsmith-engine-agents-fine-tunin… -
LangChain reposted
Nathan Tarbert
@nathan_tarbert
@LangChain Interrupt26 in NYC was amazing! 🗽
The energy of a sold out venue of agent builders from enterprises all over the world was so inspiring. @craigirwin and I spent the day demoing @CopilotKit and answering questions.
The favorite topic: how your app can have automatic learning built in 🤯 -
LangChain
@LangChain
In case you missed our Jev-inar this week with @sydneyrunkle, @allietheicon, and @huntlovell!
Sydney Runkle
@sydneyrunkle
the biggest lesson that i took away from our jev webinar w/ @allietheicon is that jev actually enables REAL TIME inference
this is because jev is fast and cheap. you can:
- do content filtering on a stream in real time
- add guardrails that don't really impact ux
- monitor (and steer) agent trajectory w/o prohibitive cost
there's a lot more to say here but i find these examples in particular very compelling
curious what others are doing in real time w/ jev!
webinar here icymi: events.langchain.com/on-demand/973158a3-ed3b-4854-8197-0f… -
LangChain
@LangChain
Join our fall AMA series for practical walkthroughs of the latest LangSmith capabilities:
📍 9/30 - Improving Agents w/ Tuned Evaluators
📍10/7 - Evaluate Agent Behavior w/ Trajectories
📍10/21 - Build and Deploy Deep Agents w/ Managed Infrastructure
events.langchain.com/fall-product-series/ -
LangChain reposted
George Pickett
@georgepickett
I built a starter app for @LangChain's new Managed Deep Agents product
Allows you to perform due diligence on @Shopify and get a cited research report
Langchain is moving fast - and the product is super easy to use.
Give it a try!This media couldn't be displayed. See it on the original post.
Parallel
@p0
Managed Deep Agents from @LangChain launched today with built-in web search powered by Parallel.
Add one MCP server to your tools file. LangSmith manages the credentials, runs the calls, and traces every search: no extra account or API key.
Bonus: It’s free during the public beta of Managed Deep Agents! -
LangChain reposted
Parag Agrawal
@paraga
Parallel is now built into @LangChain managed agents.
Julia Schottenstein
@j_schottenstein
Excited to partner with @p0 on Managed Deep Agents! You now have parallel built into your agents so your agents search the web fast and efficiently!
@travers00 @hwchase17 @VictorMoreira16 -
LangChain reposted
-
LangChain
@LangChain
And that’s a wrap on Interrupt NYC.
See you in London! -
LangChain
@LangChain
The talks may be over, but the Interrupt Reception sponsored by @tavilyai has just begun!
This media couldn't be displayed. See it on the original post. -
LangChain reposted
Ankush Gola
@ankush_gola11
Try out LangSmith Fine-Tuning today with our dedicated CLI smithtune github.com/langchain-ai/smithtune
A huge part of post training is finding and processing the right data. Our CLI handles pulling, selecting, processing trajectories from LangSmith, training on @FireworksAI_HQ and @baseten, then evaling with LangSmith.
Let me know what you think!
LangChain
@LangChain
Introducing LangSmith Fine-Tuning and the smithtune CLI.
LangSmith now handles the entire fine-tuning process. Use your traces to train specialized models that cut cost and latency.
Now in Public Beta. www.langchain.com/blog/langsmith-fine-tuning -
LangChain
@LangChain
Final sessions of Interrupt NYC
🎙️ Fireside Chat w/ @RogoAI CEO & Co-Founder @gabestengel
🎙️ Evaluating Agentic AI at Scale with @linkedin -
LangChain reposted
nathan drezner
@ndrezn
🔢 at @LangChain we've been putting a lot of thought into safe storage of user memories.
🔐 We wanted a secure way for a distributed agent to manage memory without it leaking outside of its intended context.
Here's a quick overview of the system and how to add user and agent memory to your managed deep agent - it's a single line of Python or TypeScript -
LangChain
@LangChain
Before Interrupt wraps up, stop by the LangChain Agent Arena!
🕹️ Play 30 seconds of Tetris to train your style
👉 Choose the model that plays it for you
👀 Watch it battle the reigning champion -
LangChain
@LangChain
A few things you can build:
1️⃣ Annotation queues: Review and label one (or many) runs at a time
2️⃣ Experiment comparisons: Compare outputs and metrics across experiments side by side
3️⃣ Trace reviews: Build trace or thread history views to inspect app behavior
LangChain
@LangChain
Introducing LangSmith Custom Apps
Create any interface from your agent data with a prompt. If you can think it, LangSmith can build it.
Now GA. www.langchain.com/blog/langsmith-custom-appsThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Git Maxd
@GitMaxd
Ok, this is insanely cool
@langchain is launching a model-tuning service w/@FireworksAI_HQ & @baseten
You can fine-tune your own custom model with the LangSmith Traces you already have - then compare & deploy! 👀
Hello SmithTune 🫡
LangChain
@LangChain
Introducing LangSmith Fine-Tuning and the smithtune CLI.
LangSmith now handles the entire fine-tuning process. Use your traces to train specialized models that cut cost and latency.
Now in Public Beta. www.langchain.com/blog/langsmith-fine-tuning -
LangChain reposted
Fireworks
@FireworksAI_HQ
Proud to be a launch training partner for @LangChain's LangSmith Fine-Tuning and smithtune.
LangSmith traces → managed SFT on Fireworks → eval → deploy. One CLI, no training infra.
Get a Fireworks API key: fireworks.ai/api-keys
Then run smithtune: github.com/langchain-ai/smithtune
LangChain
@LangChain
Introducing LangSmith Fine-Tuning and the smithtune CLI.
LangSmith now handles the entire fine-tuning process. Use your traces to train specialized models that cut cost and latency.
Now in Public Beta. www.langchain.com/blog/langsmith-fine-tuning -
LangChain
@LangChain
Another awesome feature in Managed Deep Agents 0.8
nathan drezner
@ndrezn
Managed Deep Agents now supports user memories.
You can control where memories are stored, and memories that are created from private contexts are never shared to other users 🔐
Context Hub acts as the management layer to securely store agent memories and make general agent memories inspectable for developers, while keeping user memories private.
Very proud of the team for this feature - durable user memory is a very tricky problem and we're really excited about the direction we're taking here. -
LangChain
@LangChain
Managed Deep Agents 0.8 makes the Slack experience even better!
Slack
@SlackHQ
Smarter permissions, seamless file drops, and built-in search. This is shaping up to be a beautiful workflow. 🎨
-
LangChain reposted
Viv
@Vtrivedy10
LangSmith Fine-Tuning 🚀
launched in partnership with our friends @baseten and @FireworksAI_HQ - try it out 🤝
one of AI’s most interesting & difficult problems today is a Data Problem
turning raw trajectories into Environments & curated trajectories to improve your agents
the goal of LangChain Labs is to help every team continuously turn their valuable trace data into better agents via Post-Training & Harness Eng
- trajectories help teams understand agent behavior at scale
- but preparing and mining raw data is really hard today, agents are not good at data curation today
- we want to help you drive this end to end flow from raw data to better agent with post-training
reach out if this is interesting and try it out 🫡
LangChain
@LangChain
Introducing LangSmith Fine-Tuning and the smithtune CLI.
LangSmith now handles the entire fine-tuning process. Use your traces to train specialized models that cut cost and latency.
Now in Public Beta. www.langchain.com/blog/langsmith-fine-tuning -
LangChain reposted
Baseten
@baseten
LangSmith Fine-Tuning from @LangChain is in public beta today, and its open-source CLI, smithtune, trains on Baseten Loops.
It turns your LangSmith traces into a fine-tuning dataset and can deploy the finished checkpoint to Baseten with one command.
Congrats on the launch!
LangChain
@LangChain
Introducing LangSmith Fine-Tuning and the smithtune CLI.
LangSmith now handles the entire fine-tuning process. Use your traces to train specialized models that cut cost and latency.
Now in Public Beta. www.langchain.com/blog/langsmith-fine-tuning -
LangChain reposted
Sydney Runkle
@sydneyrunkle
Jev + decision models featured in the @LangChain interrupt keynote!!
-
LangChain
@LangChain
Introducing LangSmith Custom Apps
Create any interface from your agent data with a prompt. If you can think it, LangSmith can build it.
Now GA. www.langchain.com/blog/langsmith-custom-appsThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Up next at Interrupt
🎙️ Fireside Chat w/ @OpenRouter CEO & Co-Founder @alexatallah
🎙️ Building Engine: How we approach agent improvement w/ @bentannyhill -
LangChain reposted
Ankush Gola
@ankush_gola11
Trajectories make it easier than ever to analyze long agent runs. They are now a first-class citizen within LangSmith.
We also spent a lot of time ensuring our trajectory format handles things like deferred tool calling and context compaction well.
LangChain
@LangChain
Introducing Trajectories in LangSmith.
A view of agent behavior that everyone on the team can understand at a glance. www.langchain.com/blog/langsmith-trajectories-tracingThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Interrupt NYC Swag spotlight:
⌨️ A clicky fidget keycap
❄️ LangChain Sweatshirt
🖊️ Pens -
LangChain reposted
Jake Broekhuizen
@jakebroekhuizen
Stoked to launch 𝚜𝚖𝚒𝚝𝚑𝚝𝚞𝚗𝚎 in public beta!
Post-training your agent's policy has never been easier, with training APIs from @baseten and @FireworksAI_HQ becoming increasingly accessible. But sourcing high-quality data to use them effectively remains a challenge.
LangSmith processes hundreds of millions of traces a day, making it a rich system of record for your agent’s behavior.
𝚜𝚖𝚒𝚝𝚑𝚝𝚞𝚗𝚎 turns those trajectories into the data you need for supervised fine-tuning, with an end-to-end pipeline for curation, preparation, training and deployment for models you want to SFT for specific tasks.
We care deeply about giving teams ways to own their own intelligence, and as the open model frontier continues to advance, it’s becoming increasingly clear that post-training on your own agent data will become a key part of that.
Excited to get this into people's hands!
LangChain
@LangChain
Introducing LangSmith Fine-Tuning and the smithtune CLI.
LangSmith now handles the entire fine-tuning process. Use your traces to train specialized models that cut cost and latency.
Now in Public Beta. www.langchain.com/blog/langsmith-fine-tuning -
LangChain
@LangChain
The LangSmith Engine that could (be won)
LangChain
@LangChain
At Interrupt?
🕹️ Try your hand at Agent Claw
🚂 Play for a LangSmith EngineThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Julia Schottenstein
@j_schottenstein
Excited to announce LangSmith Fine-Tuning in partnership with @FireworksAI_HQ @baseten. Take your LangSmith data and post train open weight models seamlessly. Pumped about this one!
-
LangChain
@LangChain
Introducing LangSmith Fine-Tuning and the smithtune CLI.
LangSmith now handles the entire fine-tuning process. Use your traces to train specialized models that cut cost and latency.
Now in Public Beta. www.langchain.com/blog/langsmith-fine-tuning
LangChain
@LangChain
1️⃣ Identify + export trajectories from LangSmith
2️⃣ Transform traces into SFT-ready training examples
3️⃣ Fine-tune open models w/ @baseten + @fireworks_AI
4️⃣ Evaluate in LangSmith
All in one simple workflow. -
LangChain reposted
Rob Bailey
@RMB
Amazing to see so many AI builders in NYC :)
-
LangChain reposted
Caspar
@caspar_br
𝚖𝚊𝚗𝚊𝚐𝚎𝚍-𝚍𝚎𝚎𝚙𝚊𝚐𝚎𝚗𝚝𝚜 0.8 is here and I think it is the best thing since sliced bread
things we shipped:
🧠 agent and user memory with access policies
🌐 connect webhooks with the new HTTP channel
✋ improved human-in-the-loop support
💬 Slack channels now support images and files
🔑 improved OAuth connection options
🔒 authenticate sandbox requests via a secure proxy
🔎 search the web with Parallel (thanks @p0)
📁 a runtime sandbox files API
released for js and py! langch.in/mda -
LangChain
@LangChain
What is a trajectory?
A conversational view of an agent session that pulls together: ✅ Human messages ✅ AI messages ✅ Tool messages ✅ System prompts …across the main agent + any subagents, then shows them in the order they first appeared.
LangChain
@LangChain
Introducing Trajectories in LangSmith.
A view of agent behavior that everyone on the team can understand at a glance. www.langchain.com/blog/langsmith-trajectories-tracingThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
After lunch, here’s what’s up next at Interrupt:
🎙️How @airbnb built Agentic AI on LangGraph
🎙️Deploying mission-critical agents to production w/ @VictorMoreira16
🎙️LangSmith for Mental Health AI w/ @SpringHealth -
LangChain
@LangChain
Introducing Trajectories in LangSmith.
A view of agent behavior that everyone on the team can understand at a glance. www.langchain.com/blog/langsmith-trajectories-tracingThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Special guests at the LangChain Demo Station: @nvidia!
Stop by and say hello during lunch! -
LangChain reposted
Ipshita Agarwal
@agarwalipshita
@p0 natively supported in @LangChain ✨
LangChain
@LangChain
Introducing Managed Deep Agents 0.8
✅ User-owned credentials + memory let agents work off correct permissions & user-specific context.
✅ HTTP channels for bringing agents into internal tools
✅ Native @slackhq file transfers
✅ Built-in web search via @p0
www.langchain.com/blog/langsmith-managed-deep-agents-what… -
LangChain reposted
nathan drezner
@ndrezn
Managed Deep Agents 0.8 launch today @ Interrupt! Check out the keynote if you haven't yet.
There's a ton of great features in this release that take MDAs to the next level for Enterprise -
- 🔐 Durable, secure user memory
- 🗄 Sandbox file API, so your agents can easily move files in and out of the sandbox
- 🌐 HTTP channels enable custom service triggers for your agent
- and more
I'll be sharing more over the course of the day and we have some great material to share about the feature set so watch this space -
LangChain
@LangChain
And it’s time for a break!
Grab lunch before we kick off the next round of sessions + workshops! -
LangChain reposted
damianztone
@damianztone
Good stuff @LangChain
-
LangChain reposted
Julia Schottenstein
@j_schottenstein
backstage with @hwchase17 @bernhardsson at @LangChain Interrupt 🤘
-
LangChain
@LangChain
Introducing Managed Deep Agents 0.8
✅ User-owned credentials + memory let agents work off correct permissions & user-specific context.
✅ HTTP channels for bringing agents into internal tools
✅ Native @slackhq file transfers
✅ Built-in web search via @p0
www.langchain.com/blog/langsmith-managed-deep-agents-what… -
LangChain
@LangChain
At Interrupt?
🕹️ Try your hand at Agent Claw
🚂 Play for a LangSmith EngineThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
New in v2:
✅ Red Teaming: Engine identifies issues in your agent before they surface in production
✅ Validated Fixes: Engine tests its proposed fixes to confirm they work before showing them to users
✅ Improved Issue Identification: Engine now surfaces inefficient agent work and trends in key metrics like cost and latency
LangChain
@LangChain
Just announced at Interrupt: LangSmith Engine v2.
With Engine v2, you can proactively spot agent issues and resolve them before your users see them.
Your agents get better. You get back to building.
www.langchain.com/blog/langsmith-engine-v2-redteamThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Next on the Interrupt main stage:
🎙️Building Jarvis: @jpmorgan’s Agent Building Assembly Line
🎙️How @useapolloio rearchitected a production GTM assistant w/ Deep Agents
🎙️Scaling Custom Agentic AI Across @Omnicom w/ @webershandwick + @focused_dot_io -
LangChain
@LangChain
Just announced at Interrupt: LangSmith Engine v2.
With Engine v2, you can proactively spot agent issues and resolve them before your users see them.
Your agents get better. You get back to building.
www.langchain.com/blog/langsmith-engine-v2-redteamThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Happening now: @modal Fireside Chat w/ CEO + Co-Founder @bernhardsson
-
LangChain reposted
Julia Schottenstein
@j_schottenstein
@bernhardsson @hwchase17 at Interrupt! Two of my favorites
@modal @LangChain -
LangChain
@LangChain
In case you missed @hwchase17’s Interrupt Keynote:
✅ LangSmith Engine v2
✅ Managed Deep Agents 0.8
✅ LangSmith Trajectories
✅ LangSmith Fine-Tuning
✅ Custom Apps
www.langchain.com/blog -
LangChain reposted
Julia Schottenstein
@j_schottenstein
Excited to partner with @p0 on Managed Deep Agents! You now have parallel built into your agents so your agents search the web fast and efficiently!
@travers00 @hwchase17 @VictorMoreira16 -
LangChain
@LangChain
Packed house for @hwchase17’s opening keynote!
-
LangChain reposted
Julia Schottenstein
@j_schottenstein
Interrupt NYC!!!!! @hwchase17 🔥
-
LangChain
@LangChain
Join the opening keynote livestream: interrupt.langchain.com/nyc/livestream
This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
🔴 Going live in 15!!!
LangChain
@LangChain
Want to see what’s next for agents?
We’re livestreaming @hwchase17’s keynote at 9:30AM ET!
interrupt.langchain.com/nyc/livestream -
LangChain
@LangChain
It wouldn’t be an agent conference without humans in the loop.
This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
At Interrupt? Grab some breakfast + coffee from our sponsors at @focused_dot_io + @FactoryAI before we kick off!
-
LangChain
@LangChain
Want to see what’s next for agents?
We’re livestreaming @hwchase17’s keynote at 9:30AM ET!
interrupt.langchain.com/nyc/livestream -
LangChain
@LangChain
Today’s the day!
Interrupt has arrived in NYC 🗽 -
LangChain
@LangChain
Next week, the LangSmith Roadshow heads to Los Angeles!
✅ Learn everything you need to know about Deep Agents and how to improve your agents with LangSmith Engine
✅ Hear how @ServiceTitan builds agents and what they’ve learned along the way
events.langchain.com/LangSmithRoadshow/LosAngeles/ -
LangChain reposted
Sydney Runkle
@sydneyrunkle
quick explainer on what it means to be a "system one" model (like jev from @typesafeai)
-
LangChain reposted
Harrison Chase
@hwchase17
Proactive agents run in the background, schedules are (one way) how you achieve that!
LangChain
@LangChain
You can add schedules to Managed Deep Agents.
1️⃣ Add a schedule file to schedules/ with a cron expression + a prompt
2️⃣ Deploy
...And the agent runs itself
💻 Docs: docs.langchain.com/langsmith/python/managed-deep-agents-s…
⏯️ Demo from @ndrezn ⤵️This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
You can add schedules to Managed Deep Agents.
1️⃣ Add a schedule file to schedules/ with a cron expression + a prompt
2️⃣ Deploy
...And the agent runs itself
💻 Docs: docs.langchain.com/langsmith/python/managed-deep-agents-s…
⏯️ Demo from @ndrezn ⤵️This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
🗽 Tomorrow! Interrupt NYC!
Austin Vance
@austinbv
What's better than christmas? Three @LangChain Interrupts in a year!
See you in NYC tomorrow -
LangChain
@LangChain
🧵 @colifran_ on everything new in OpenWiki 0.6.0
Colin Francis
@colifran_
openwiki 0.6.0 is out 🎉
here's what's new:
👉 group wikis into shared workspaces with openwiki link
👉 generate code wikis faster with new concurrency controls
👉 use openwiki with @antigravity and oh my pi
🧵 1/4 -
LangChain
@LangChain
We're releasing new products on Thursday, and we're showing them first on the Interrupt keynote livestream.
@hwchase17 will walk through what's working for agents in production and where we think they're headed, then show what we've been building.
Save your spot ⤵️ -
LangChain reposted
Harrison Chase
@hwchase17
🖼️ Improved UI for Jev models in LangSmith
agents are going to complex systems of many llms calls
many llm calls will be jev-like (decision models)
we've updated tracing view in LangSmith to display these types of calls (state, questions, choices, output) more clearly! -
LangChain
@LangChain
On September 30th, learn about the key elements of online evals and see how teams can use Tuned Evaluators in LangSmith to automatically analyze agent interactions and attach useful feedback to production traces. events.langchain.com/register/fall-series/tuned-evaluators/
-
LangChain reposted
Sydney Runkle
@sydneyrunkle
this agent
1. browses the web with stagehand from @browserbase
2. prepares payment with the link CLI from @stripe
3. asks for approval before the charge
4. places the order, ready for pickup!
official @LangChain / link integration coming soon! -
LangChain
@LangChain
Join @AWScloud, @NVIDIA, and LangChain for an in-person session focused on how teams can build flexible infrastructure for developing and scaling agentic AI workloads.
We’ll explore how Amazon EKS, NVIDIA technologies, and LangChain can come together to give teams more flexibility as they move from experimentation toward production agent systems.
Register here: aws-experience.com/apj/smb/e/0700a/architecting-flexible-… -
LangChain
@LangChain
LangSmith now supports decision models including Jev and SemIf, giving you visibility into every step. Now, you can:
✅ Debug faster
✅ Understand model behavior
🚀 Confidently improve your apps
Try it today -> smith.langchain.com -
LangChain reposted
Sydney Runkle
@sydneyrunkle
jev might be a new type of model, but one thing that doesn't change is the need for observability
you still need to answer questions like "what decisions is jev making" and "are those decisions good for my use case"?
langsmith helps you with that, this new UI is very clean!
Harrison Chase
@hwchase17
🖼️ Improved UI for Jev models in LangSmith
agents are going to complex systems of many llms calls
many llm calls will be jev-like (decision models)
we've updated tracing view in LangSmith to display these types of calls (state, questions, choices, output) more clearly! -
LangChain
@LangChain
Healthcare organizations are using agents to transform patient care across workflows. But one persistent bottleneck is that clinical expertise is a scarce resource, and these experts are typically ones that must vet the accuracy of these agents.
LangSmith is helping transform expert review into durable + reusable evaluations, compounding the value of human judgment.
Read on to see real-world examples of this loop in action. -
LangChain
@LangChain
With Managed Deep Agents, you can build agents that ship GitHub PRs from Slack with just a few lines of code.
We built one called Patch
💻 Tag it in a thread
💬 Describe the feature
…and it opens a PR with the diff, ready to review.
docs.langchain.com/langsmith/python/managed-deep-agents-o…This media couldn't be displayed. See it on the original post. -
LangChain reposted
Sydney Runkle
@sydneyrunkle
ICYMI -- quick video walkthrough of jev, system one models, and how to build them into your agent harness!
Sydney Runkle
@sydneyrunkle
here's a quick breakdown of @typesafeai's new jev model: what it is, and how to plug it into your agents!
This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Today at 11am PT / 2pm ET.
Join us + @typesafeai for a live session on building an agent harness with Jev!
RSVP: events.langchain.com/webinar/building-a-harness-with-jev/ -
LangChain
@LangChain
.@Clay evaluates every agent on three things:
✅ Quality
✅ Throughput
✅ Cost
@jeffbarg on how LangSmith's evals measure quality and benchmark it against cost and throughput.This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
We’re one week away from the LangSmith Roadshow in Boston!
✅ Get deep dives on LangSmith Engine + Deep Agents
✅ Hear how @censysio uses LangSmith to build, debug, and monitor agents running in production
RSVP: events.langchain.com/LangSmithRoadshow/Boston/ -
LangChain
@LangChain
Happening tomorrow! @LangChain_OSS team x @TypeSafeAI livestream.
Come learn about Jev with @sydneyrunkle, @huntlovell, and @allietheicon.
RSVP: events.langchain.com/webinar/building-a-harness-with-jev/ -
LangChain reposted
Harrison Chase
@hwchase17
try out jev in langsmith today! can use it to score traces cheap and accurately
LangChain
@LangChain
Jev-as-a-judge is now available in LangSmith.
✅ Score every production trace instead of a sample.
✅ Check more criteria per trace without the cost climbing.
✅ Catch safety or security issues fast enough to trigger an automated response.
Give it a try and let us know what you think!
www.langchain.com/blog/jev-is-now-available-in-langsmith-… -
LangChain
@LangChain
Give it a try → docs.langchain.com/oss/python/integrations/providers/type…
nathan drezner
@ndrezn
If you're using managed deep agents in LangChain, take our @typesafeai Jev middleware for model routing for a spin.
Define your routes and Jev handles model selection. -
LangChain reposted
nathan drezner
@ndrezn
If you're using managed deep agents in LangChain, take our @typesafeai Jev middleware for model routing for a spin.
Define your routes and Jev handles model selection. -
LangChain
@LangChain
Jev-as-a-judge is now available in LangSmith.
✅ Score every production trace instead of a sample.
✅ Check more criteria per trace without the cost climbing.
✅ Catch safety or security issues fast enough to trigger an automated response.
Give it a try and let us know what you think!
www.langchain.com/blog/jev-is-now-available-in-langsmith-…
LangChain
@LangChain
We tested Jev against GPT-5.6 Luna, GPT-5.6 Terra, and Claude Sonnet 4.6 as agent-eval judges.
✅ Jev matched a human reviewer on every call, with up to 913x lower variance, at 0.44s per call vs 2.16-2.83s for the LLMs.
✅ Cost for the full run: $0.34 with Jev. $28.17 with Claude Sonnet 4.6. -
LangChain
@LangChain
We're making SemIf, an open-source decision model, free for the next week in LangSmith LLM Gateway.
Jacob Lee
@Hacubu
🆓 Free Open-Source Jev 🤝 LangSmith Gateway
Decision models are becoming a first-class part of the LangSmith Gateway!
We're serving SemIf, an open-source decision model, free for the next week.
Compatible with the TypeSafe SDK - just change a string to try it!
Docs in 🧵
LangChain
@LangChain
Check out the docs: docs.langchain.com/langsmith/llm-gateway-decision-models
-
LangChain
@LangChain
In <10 minutes, @sydneyrunkle walks through everything you need to know about Jev.
Sydney Runkle
@sydneyrunkle
here's a quick breakdown of @typesafeai's new jev model: what it is, and how to plug it into your agents!
This media couldn't be displayed. See it on the original post. -
LangChain reposted
Jacob Lee
@Hacubu
🆓 Free Open-Source Jev 🤝 LangSmith Gateway
Decision models are becoming a first-class part of the LangSmith Gateway!
We're serving SemIf, an open-source decision model, free for the next week.
Compatible with the TypeSafe SDK - just change a string to try it!
Docs in 🧵 -
LangChain
@LangChain
A simple way to think about how Jev works vs traditional LLMs
Sydney Runkle
@sydneyrunkle
why is jev called a "system one" model?
this actually comes from @kahneman_daniel's book "thinking, fast and slow" that outlines two ways of thinking:
system one: fast, cheap, almost automatic (jev)
system two: slow, deliberate, and analytical (llms) -
LangChain
@LangChain
.@elastic is sponsoring Interrupt NYC!
Stop by their booth during the event and see how you can power insights and outcomes with The Elasticsearch Platform.
LangChain
@LangChain
Learn more about Interrupt, The Agent Conference by LangChain and RSVP: interrupt.langchain.com
-
LangChain
@LangChain
As more global teams move voice agents from experimentation into production, evaluating whether a call “worked” is no longer enough.
A voice agent can follow instructions correctly and still create a poor user experience. Or it can complete a task without achieving the intended business outcome.
In our upcoming webinar, we’ll break evaluation into three dimensions:
- Execution: Did the agent follow its instructions and use tools correctly?
- Outcomes: Did the interaction actually achieve its goal?
- Experience: Was the conversation fast, clear, and natural?
Join us to learn how to build a more complete evaluation approach for production voice agents.
Register: events.langchain.com/register/emea-evaluating-voice-agents/ -
LangChain
@LangChain
We're hosting a hot topic livestream on Jev with
@typesafeai on Tuesday featuring @sydneyrunkle @allietheicon @huntlovell
Come join!
events.langchain.com/webinar/building-a-harness-with-jev/ -
LangChain reposted
Sydney Runkle
@sydneyrunkle
let the building begin!!
a simple guide on what makes jev special, how to use it w/ your agents, and a few model use cases: x.com/sydneyrunkle/status/2100754364545761643
TypeSafe AI
@typesafeai
Jev is now available to everyone. No waitlist.
Start using it here: console.typesafe.ai -
LangChain reposted
Sydney Runkle
@sydneyrunkle
jev as a Judge proves to be a cheaper and more precise alternative to LLM as a judge for online evals.
great guide from Sean and Daniel on why Jev is great for evals, an experiment vs other LLMs, and how to try this out for your agents!
LangChain
@LangChain
We tested Jev against LLM judges on accuracy, repeatability, latency, and cost to see whether System One models could offer a new approach to agent evaluation. x.com/i/article/2101448785255907328
-
LangChain
@LangChain
We tested Jev against LLM judges on accuracy, repeatability, latency, and cost to see whether System One models could offer a new approach to agent evaluation. x.com/i/article/2101448785255907328
-
LangChain reposted
Jess Ou
@JessicaOu
LangChain is hosting a hot topic livestream on Jev with @typesafeai on Tuesday with @sydneyrunkle @allietheicon
Come join! events.langchain.com/webinar/building-a-harness-with-jev/ -
LangChain reposted
Sydney Runkle
@sydneyrunkle
jev is so exciting! so you started using jev, now you need to know
1. what decisions is jev making? (observability)
2. is it making the right ones for my application? (evals)
3. how can i help jev make better decisions? (engine)
and for obs+evals+engine, you need langsmith -
LangChain reposted
Git Maxd
@GitMaxd
Big fan of how @LangChain is doing these insta-webinars
The news is a few days old, @typesafeai is integrated and documented with example code in LangChains Docs 👇
And they are doing a webinar on the topic … in the same seven days 🦜
LangChain
@LangChain
Hot topic livestream: Learn about Jev
A buzzy new model Jev, by @typesafeai, reports up to 200x faster inference and 400x lower cost than comparable LLMs on classification tasks. It's aiming to solve the problem of slow and costly agent loops.
Next Tuesday, we’re bringing @sydneyrunkle from our engineering team at @LangChain, for a live convo on how TypeSafe AI’s model makes fast, structured decisions, where it fits in the agent loop, and how to build a harness with Jev.
Join here: events.langchain.com/webinar/building-a-harness-with-jev/ -
LangChain reposted
Grant Miller
@AIGuide_
if you're confused about Jev this is 100% worth your time
in fact you should study it
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
Hot topic livestream: Learn about Jev
A buzzy new model Jev, by @typesafeai, reports up to 200x faster inference and 400x lower cost than comparable LLMs on classification tasks. It's aiming to solve the problem of slow and costly agent loops.
Next Tuesday, we’re bringing @sydneyrunkle from our engineering team at @LangChain, for a live convo on how TypeSafe AI’s model makes fast, structured decisions, where it fits in the agent loop, and how to build a harness with Jev.
Join here: events.langchain.com/webinar/building-a-harness-with-jev/ -
LangChain reposted
Harrison Chase
@hwchase17
browser use is the best application of Jev ive seen yet
nathan drezner
@ndrezn
Browser use w/ @LangChain + @typesafeai's Jev! Really fun to build.
... I found it's excellent at playing the Wikipedia Game. (But, it's also great at "folding laundry" type tasks, like finding cheap flights.) -
LangChain reposted
nathan drezner
@ndrezn
Browser use w/ @LangChain + @typesafeai's Jev! Really fun to build.
... I found it's excellent at playing the Wikipedia Game. (But, it's also great at "folding laundry" type tasks, like finding cheap flights.) -
LangChain
@LangChain
We're demoing our Paid Media Agent live on September 23rd.
We’ll cover the architecture, how it works, what we’ve learned in production, and answer your questions along the way!
events.langchain.com/gtm-engineering-live-series/paid-med… -
LangChain reposted
LangChain
@LangChain
LangChain 🤝 Jev!
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
Jev explained in <20 seconds by @sydneyrunkle
Sydney Runkle
@sydneyrunkle
instead of generating text, jev from @typesafeai generates structured output
this makes it great for classification tasks like model routing, tool selection/search, and guardrails of many forms!
it's also ridiculously fast and cheap compared to LLMs doing the same tasks -
LangChain
@LangChain
.@PodiumHQ migrated off their own agent runtime and onto LangSmith's.
Walker Ward on why building your own is rarely worth it anymore.This media couldn't be displayed. See it on the original post.
LangChain
@LangChain
Watch the full conversation: www.youtube.com/watch?v=J77ro1AJGa0
-
LangChain reposted
Sydney Runkle
@sydneyrunkle
right now, a lot of agents perform classification tasks / steps
if you can replace those steps with a model that's optimized for classification, you can massively reduce latency and cost
i think jev is re-kindling the "graph engineering" phase of agent design
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
.@redisinc is sponsoring Interrupt NYC!
Stop by their booth during the event and go hands-on with their data platform, built for caching, AI, and real-time apps.
LangChain
@LangChain
Learn more about Interrupt, The Agent Conference by LangChain and RSVP for our livestream: interrupt.langchain.com
-
LangChain
@LangChain
Teams building reliable agents need the right harness for planning, tool use, context, and long-running tasks.
Join us for our Agent Harnesses & Deep Agents 101 Virtual Workshop, where we’ll break down the patterns behind Deep Agents and show how to build agents that can take on more complex, multi-step work.
🌍 EMEA: events.langchain.com/register/emea-agent-harnesses-and-de…
🌏 APAC + ANZ: events.langchain.com/register/apac-anz-agent-harnesses-an… -
LangChain
@LangChain
LangChain 🤝 Jev!
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
LangSmith for Startups Spotlight: @enzo_health
Enzo uses LangSmith to trace and evaluate its clinical AI pipelines, from intake extraction to visit documentation, letting the team debug individual patient encounters and catch regressions before they ship, all within the compliance requirements of HIPAA-regulated healthcare.
Enzo is hiring engineers who want to ship AI that clinicians actually use. Check out open roles at enzo.health/careers. -
LangChain
@LangChain
Included Health built Dot, a federated multi-agent healthcare guide, on LangGraph + Deep Agents, backed by LangSmith for clinical oversight.
Multi-turn simulation evals got 4 teams onto Deep Agents in under 2 weeks, no significant regressions.
✅ 75% more chat engagement
✅ 99%+ high-risk detection
This is their story:
www.langchain.com/blog/how-included-health-built-federate… -
LangChain
@LangChain
Deep Life Sci can access clinical trials records from:
✅ 600k+ registered studies on clinicaltrials.gov
✅ 29m paper abstracts through PubMed
✅ 12m full-text articles on PubMed Central
Able to review hundreds of docs at once via sub-agents
Every agent equipped with a LangSmith Sandbox
Get started: github.com/langchain-samples/deep-life-sci
LangChain
@LangChain
Meet Deep Life Sci: An open source agentic assistant created specifically for clinical and lab scientists, built on Deep Agents.
www.langchain.com/blog/agent-harness-life-sciences -
LangChain
@LangChain
New in OpenWiki v0.5.2.
@IBM Bob Shell integration.
Colin Francis
@colifran_
openwiki welcomes bob shell as the second coding agent integration in openwiki v0.5.2!
get started with:
1) npm install -g openwiki@latest
2) openwiki integrations install bob
we're at 6 coding agent integrations and growing. let us know what you want to see next! -
LangChain
@LangChain
Great deep dive into our Paid Media Agent from @amal_irgashev ⤵️
amal
@amal_irgashev
Our paid media agent at @LangChain has 200+ tools, a 19-page company wiki and its own computer with 8 libraries for data analysis and reporting
It writes code, analyzes campaigns, generates branded reports and helps our marketing team prepare campaign changes and new launches in Slack
Here are 6 things we learned about context engineering, tools and runtime design while building it🧵 -
LangChain
@LangChain
Meet Deep Life Sci: An open source agentic assistant created specifically for clinical and lab scientists, built on Deep Agents.
www.langchain.com/blog/agent-harness-life-sciences -
LangChain reposted
Arjun Natarajan
@arjunnatarajan_
this has been a fun one to work on! agent observability means nothing if you cant easily find the traces you're looking for. we rebuilt filtering from the ground up and are excited for you all to try it!
LangChain
@LangChain
Filtering agent traces in LangSmith just got faster 🔎
We rebuilt the filtering experience. Now, it's easier than ever to:
🧩 Build precise queries to narrow down your runs
🎯 Find the exact runs you need
🧭 Understand why a result matched
Start tracing -> smith.langchain.comThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Filtering agent traces in LangSmith just got faster 🔎
We rebuilt the filtering experience. Now, it's easier than ever to:
🧩 Build precise queries to narrow down your runs
🎯 Find the exact runs you need
🧭 Understand why a result matched
Start tracing -> smith.langchain.comThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Next stops on the LangSmith Roadshow?
📍 9/29: Boston
📍 10/1: Los Angeles
If you want a hands-on look at how to build and improve agents with agents, you won’t want to miss these half-day events.
events.langchain.com/LangSmithRoadshow/ -
LangChain
@LangChain
See how how @SchneiderElec, @Vodafone, and @mondaydotcom are approaching production AI with stronger systems for observability, evaluation, orchestration, governance, and scale.
See the patterns emerging across leading teams in the region: info.langchain.com/guide/scaling-agents-in-europe-and-the… -
LangChain reposted
amal
@amal_irgashev
Introducing Paid Media Agent
An open-source agent for paid campaign analysis, reporting and management, built with Managed Deep Agents
✓ Compare spend, conversions and cost per lead across channels
✓ Get weekly performance reports in Slack
✓ Turn briefs into campaign drafts, or prepare changes to budgets, keywords and ad copy
Here’s what you can do with it, and how to make it your own 🧵This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
RT @amal_irgashev: Excited to share more of the Paid Media Agent we built with @LangChain
It runs our weekly campaign analysis and report…
LangChain
@LangChain
A quick TL;DR on our Paid Media Agent:
✅ A Deep Agent that lives in Slack.
✅ Reads spend, clicks, and conversions across 6 ad platforms + our data warehouse every Monday.
✅ Tells our team what changed and why, proposes changes, and applies once approved.This media couldn't be displayed. See it on the original post. -
LangChain reposted
-
LangChain
@LangChain
A quick TL;DR on our Paid Media Agent:
✅ A Deep Agent that lives in Slack.
✅ Reads spend, clicks, and conversions across 6 ad platforms + our data warehouse every Monday.
✅ Tells our team what changed and why, proposes changes, and applies once approved.This media couldn't be displayed. See it on the original post.
LangChain
@LangChain
Want to learn more?
💻 Full engineering breakdown: www.langchain.com/blog/paid-media-agent
🔴 We're demoing it live on Sept 23: events.langchain.com/gtm-engineering-live-series/paid-med… -
LangChain
@LangChain
🎧 @hwchase17 joined @browserbase founder @pk_iv on the latest episode of the Navigators podcast.
Great talk on:
✅ Why the best agent builders lead with the outcome they're going for + let the tools follow
✅ Why a model gateway needs to be core infrastructure, not something bolted on later
YouTube: www.youtube.com/watch?v=oP_W7GuP6OA
Spotify: open.spotify.com/episode/65ipDP5AJGWrwtqaPBHg4r?nd=1&dlsi…This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
New in OpenWiki v0.5.2.
@kirodotdev integration.
Colin Francis
@colifran_
openwiki 🤝 kiro
openwiki v0.5.2 dropped recently and we added two new coding agent integrations. the first one i wanted to showcase is kiro!
coding agent integrations make it easier than ever to get started with openwiki
1) npm install -g openwiki@latest
2) openwiki integrations install kiro
check it out ⤵️ -
LangChain
@LangChain
PSA: You can add custom middleware to your Managed Deep Agents.
With middleware, you can add behavior around model calls, tool calls, and the agent lifecycle.
Here's a quick demo.This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Clay built its agent harness on the Vercel AI SDK.
Getting complete tracing into LangSmith took about one line of code.
@jeffbarg on what LangSmith unlocked after that.
LangChain
@LangChain
Watch the full conversation: www.youtube.com/watch?v=cx6_tb6HCeY
-
LangChain reposted
Sydney Runkle
@sydneyrunkle
building an agent (model + harness) is about 2 things:
1. picking the right model for the job. this means finding the sweet spot on the cost/intelligence curve.
2. building a harness that's fit to the agent's task(s). this means the harness can get the right context to the model at any step.
here's a guide on how to build a domain specific harness: www.langchain.com/blog/how-to-build-a-custom-agent-harness -
LangChain
@LangChain
The biggest challenge facing an agent harness is context engineering, or what information an agent gets, and when.
The longer the task, the harder it is to keep the context window from overloading.
Deep Agents ships with this built in: filesystems, subagents, and skills, already wired up for context management.This media couldn't be displayed. See it on the original post.
LangChain
@LangChain
Learn more in our LangChain Academy course: academy.langchain.com/courses/foundation-introduction-to-…
-
LangChain
@LangChain
.@Baseten is sponsoring Interrupt NYC!
Stop by their booth during the event and go hands-on with the platform for high-performance inference.
LangChain
@LangChain
Learn more about Interrupt, The Agent Conference by LangChain and RSVP: interrupt.langchain.com
-
LangChain
@LangChain
It used to take @Lyft six months to ship a support agent, and nobody could tell why one would misbehave in production.
With LangGraph and LangSmith, PMs can write a prompt, push a config change, ship it in 1-2 weeks, and see what their agent did.This media couldn't be displayed. See it on the original post. -
LangChain reposted
Jake Broekhuizen
@jakebroekhuizen
Always a great time with @AEllisBloor and the Baseten crew 👑 . The Aussie SF contingent runs strong.
-
LangChain reposted
Viv
@Vtrivedy10
The future is about continuously turning your valuable data & ideas into better/faster/cheaper models -> Owning your Intelligence
the Baseten team rocks at helping us run tons of experiments to make this happen @AEllisBloor @mudithj 🤝
A lot of work in our Labs team looks like:
- sitting down with a buddy @jakebroekhuizen
- picking a problem/agent to make better, cheaper, faster
- curating data + evals/environments
- doing tons of training experiments, with continuous measurement of every tweak
- Loop (pun intended)
the infrastructure and tooling is here to allow every team to run hundreds of experiments and get the best quality/cost/latency tradeoff for their customers and use-cases
also ngl training is fun and your agents will love the experience -
LangChain
@LangChain
.@Baseten Loops helps our team train the models behind LangSmith Engine.
✅ Fine-tune open-weight models on agent traces
✅ Ship checkpoints straight to production
✅ Own our intelligence
✅ Help top teams ship great agents -
LangChain
@LangChain
🔌 Every Managed Deep Agent is an MCP server.
Let coding assistants and other agents delegate research, analysis, and more to your deployed agent as a tool.
Deploy once. Reuse across compatible MCP clients through the built-in /mcp endpoint.
Get started → -
LangChain
@LangChain
🎓 New in our LangSmith Essentials course: A capstone
We’ve added even more opportunities to apply your debugging and tracing skills, and for you to master our framework agnostic agent engineering platform for observing, evaluating, and deploying agents.
Get started at LangChain Academy today ⤵️
academy.langchain.com/courses/quickstart-langsmith-essent… -
LangChain
@LangChain
Voice agents are becoming a bigger part of customer and operational workflows globally, but getting them into production requires more than making the conversation sound natural.
Join us for a practical session focused on a practical framework for evaluating production voice agents.
We’ll cover how to assess tool use and task completion, connect conversations to real-world outcomes, and identify voice-specific issues like latency, interruptions, pacing, and conversational friction.
events.langchain.com/register/emea-evaluating-voice-agents/ -
LangChain reposted
Steve Kaliski
@stevekaliski
we love langchain!
-
LangChain
@LangChain
The OSS team is cooking on small but mighty harness optimizations 💪
Sydney Runkle
@sydneyrunkle
we just changed how deepagents read files!
our evals show that this new format reduces edit_file errors by 15% AND total input token usage by 10%
the little things really add up w/ agents; we're hard at work on these small but mighty optimizations at @LangChain -
LangChain
@LangChain
Can't join Interrupt NYC in person? We're hosting a livestream of the opening keynote on September 24th!
Tune in from 9:30-10:00 AM ET as @hwchase17 shares what's coming next for agents: what's actually working in production today, where the industry is headed, and an exclusive first look at new product releases. interrupt.langchain.com/nyc/livestream -
LangChain
@LangChain
Watch @Vtrivedy10 take @businessbarista from easy mode to god mode on all things evals.
Alex Lieberman
@businessbarista
A few of my smartest friends in AI called me a "idiot" for not deeply understanding evals.
So...I found the smartest person I know on evals & made them teach me.
@Vtrivedy10 (leads Labs at @LangChain) took me from easy mode to god mode for a 38-minute masterclass on all things evals.
Easy Mode: what an eval actually is
Definition: did the AI agent do the job correctly?
You need two building blocks:
1) Tasks. The checkable jobs you care about. Log the meeting. Draft the email. Find Acme across the right Salesforce tables.
2) Verifiers. Something that can say right or wrong after the task. A script. Another model. A human with a clear checklist.
Hard Mode: what are environments
Definition: a safe practice field for your agent to do work & for you to evaluate its performance.
Rules of thumb:
1) Never test on production. Agents will cheat because they're optimizing for the score you gave them.
2) If you're not an engineer, you still have options for running environments/evals.
- Harbor (open source primitives for tasks, verifiers, sandboxes)
- LangSmith Engine (UI for people who can judge good vs bad without living in GitHub).
- Steal a published Harbor-format eval, ask Claude Code or Codex to explain it, then tweak it for your agent.
God Mode: what is a self-improving loop
Definition: Run the agent in the real world --> turn that production behavior into evals/environments --> change the agent so failures stop happening --> repeat
Rules of thumb:
1) Turn on tracing first. Traces = receipts of every action (tool calls, Salesforce pings, web searches, dead ends).
2) Store those logs somewhere (LangSmith at org scale, or even “have the agent read its own output files” at small scale).
3) Point a second agent at the first agent’s traces to spot patterns (“always searches the wrong tables,” “multi-company asks collapse to one company”) and propose fixes overnight if your eval suite is solid.
Full episode: www.youtube.com/watch?v=zLeG-XJtbIEThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Viv
@Vtrivedy10
Alex & the Tenex team rock 🙏
Evals & Environments & Data are driving a huge chunk of the world’s AI progress from the frontier labs, to open source, to every company serving agents to real users.
At @LangChain we want to help every team own the process of turning their valuable data into evals that improve their agents over time
- Evals are basically Tasks/Tests that help you measure how an agent does on work you care about
- Evals today look different than 3 years ago (no more prompt in/out). Today’s agents do much more autonomous work and we need Environments that faithfully capture the real work they do
- Building good evals today is still hard, this is why frontier companies pay millions to acquire them across domains
- Every team looking to Own their Intelligence will need to own their Evals. Both so they can measure the progress of agents on their tasks and also because Evals are literally the thing that shapes what your agent will become.
- When you train on Evals or hill-climb them with harness-engineering, you transfer the behaviors from the eval into your agent. Garbage in garbage out - the inescapable lesson of machine learning
there’s so much to gain from helping every team own their Data to Eval pipeline, let’s cook 🚀
Alex Lieberman
@businessbarista
A few of my smartest friends in AI called me a "idiot" for not deeply understanding evals.
So...I found the smartest person I know on evals & made them teach me.
@Vtrivedy10 (leads Labs at @LangChain) took me from easy mode to god mode for a 38-minute masterclass on all things evals.
Easy Mode: what an eval actually is
Definition: did the AI agent do the job correctly?
You need two building blocks:
1) Tasks. The checkable jobs you care about. Log the meeting. Draft the email. Find Acme across the right Salesforce tables.
2) Verifiers. Something that can say right or wrong after the task. A script. Another model. A human with a clear checklist.
Hard Mode: what are environments
Definition: a safe practice field for your agent to do work & for you to evaluate its performance.
Rules of thumb:
1) Never test on production. Agents will cheat because they're optimizing for the score you gave them.
2) If you're not an engineer, you still have options for running environments/evals.
- Harbor (open source primitives for tasks, verifiers, sandboxes)
- LangSmith Engine (UI for people who can judge good vs bad without living in GitHub).
- Steal a published Harbor-format eval, ask Claude Code or Codex to explain it, then tweak it for your agent.
God Mode: what is a self-improving loop
Definition: Run the agent in the real world --> turn that production behavior into evals/environments --> change the agent so failures stop happening --> repeat
Rules of thumb:
1) Turn on tracing first. Traces = receipts of every action (tool calls, Salesforce pings, web searches, dead ends).
2) Store those logs somewhere (LangSmith at org scale, or even “have the agent read its own output files” at small scale).
3) Point a second agent at the first agent’s traces to spot patterns (“always searches the wrong tables,” “multi-company asks collapse to one company”) and propose fixes overnight if your eval suite is solid.
Full episode: www.youtube.com/watch?v=zLeG-XJtbIEThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Harrison Chase
@hwchase17
slack is where agents are doing work
managed deepagents has a very tight slack integration - check out the tutorial on it below
LangChain
@LangChain
PSA: You can connect a Managed Deep Agent to Slack.
Mentions, DMs, and thread replies all start a run. Your agent posts its response back in the same conversation.
Here’s how to get started ⤵️This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
PSA: You can connect a Managed Deep Agent to Slack.
Mentions, DMs, and thread replies all start a run. Your agent posts its response back in the same conversation.
Here’s how to get started ⤵️This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
LangSmith LLM Gateway lets you govern model access across your org.
🔒 Lock a key to permitted models, and every other model it tries to call gets blocked automatically.
See what happens when we allow one key to call Opus 5 only, then try Sonnet 5 with that same key.This media couldn't be displayed. See it on the original post.
LangChain
@LangChain
Learn more about LangSmith LLM Gateway www.langchain.com/langsmith/llm-gateway
-
LangChain
@LangChain
.@Antimetal is sponsoring Interrupt NYC!
Stop by their booth during the event and go hands-on with their autonomous system to manage your software in production.
LangChain
@LangChain
Learn more about Interrupt, The Agent Conference by LangChain and RSVP: interrupt.langchain.com
-
LangChain
@LangChain
Agents are already helping healthcare and life sciences teams accelerate research, reduce clinical documentation burden, and make complex healthcare data easier to access.
Our latest guide looks at how @MadrigalPharma, @AbridgeHQ, and Vizient are building production agent systems with stronger infrastructure for evaluation, observability, orchestration, governance, and deployment.
info.langchain.com/guide/scaling-agents-in-healthcare-and… -
LangChain
@LangChain
It wouldn't be a trip to NYC without a stop in Times Square.
@BrexHQThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Agent harnesses explained in <90 seconds.
TLDR: if you’re not working on the model, you’re working on the harness.
academy.langchain.com/courses/foundation-introduction-to-…This media couldn't be displayed. See it on the original post. -
LangChain reposted
Adam Łucek
@AdamRLucek
I've been working on Engine, the "Agent for Agent Engineering," for the last 5 months and (while I might be a little biased) I think it's pretty cool- so I wanted to talk a bit about how we've been approaching automated agent development with an agent of our own using trace data and project context 🧵
-
LangChain
@LangChain
Great thread from @AdamRLucek on LangSmith Engine!
Adam Łucek
@AdamRLucek
I've been working on Engine, the "Agent for Agent Engineering," for the last 5 months and (while I might be a little biased) I think it's pretty cool- so I wanted to talk a bit about how we've been approaching automated agent development with an agent of our own using trace data and project context 🧵
-
LangChain
@LangChain
LangSmith is @PodiumHQ's sanctioned go-to platform for every team building agents.
Walker Ward on why one shared platform beats a dozen one-off ones.This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Agents are moving quickly into production across Europe and the Middle East.
Our latest guide looks at how @SchneiderElec, @Vodafone, and @mondaydotcom are building the infrastructure and operating practices needed to scale agents across the enterprise.
Read the guide: -
LangChain reposted
Sydney Runkle
@sydneyrunkle
gooood morning SF and the @LangChain engine bus ☺️
-
LangChain
@LangChain
Credit Genie's coding agents used to guess at how a repo worked. Now they just check their OpenWiki.
LangChain
@LangChain
The loop runs without anyone remembering to run it:
New repo → added to config → OpenWiki opens a stub PR → merged, repo's onboarded.
Every night, it checks commits and opens a PR when something's changed.
A daily sweep merges pending PRs and rebuilds the portal.
LangChain
@LangChain
Full story on how Credit Genie set this up: www.langchain.com/blog/how-credit-genie-uses-openwiki-to-…
-
LangChain
@LangChain
Add skills to your agent, swap the model, tweak a tool description. How do you know you didn't break what already worked?
Managed Deep Agents ships with @harborframework.
Evals run in a fresh container every time, tracked in LangSmith.
What you need to know in 9 minutes:This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Point a coding agent at the LangSmith CLI and it can build you an online LLM-as-a-judge: define the rubric, wire it to every incoming trace, adjust the sampling rate, all from your terminal.
Here's a 5 minute tutorial.This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Manual review works at small scale. At millions of agent runs a month, it falls apart, so @Clay relies on LangSmith's online evaluators and is testing its insights product to understand agent behavior.
@jeffbarg explains.
LangChain
@LangChain
Watch the full conversation: www.youtube.com/watch?v=cx6_tb6HCeY
-
LangChain
@LangChain
.@tavilyai is sponsoring Interrupt NYC!
Stop by their booth during the event and see how you can connect your AI agents to the web.
LangChain
@LangChain
Learn more about Interrupt, The Agent Conference by LangChain and RSVP: interrupt.langchain.com
-
LangChain reposted
Sydney Runkle
@sydneyrunkle
models are really good at writing code :)
first we saw coding agents explode in usage, then we saw agents shifting to "code mode" style workflows
computer use is an excellent application of this -- agents are better at using a computer if they can navigate with code rather than a fixed set of tools
check out this article on how to build effective browser agents with deepagents and browserbase!
Kyle Jeong
@kylejeong
-
LangChain
@LangChain
The @browserbase team is working on building the best tools for agents to use the browser
It's super easy to attach these to your deepagents:
docs.langchain.com/oss/python/integrations/tools/stagehand
Kyle Jeong
@kylejeong
-
LangChain
@LangChain
NYC!
You’re invited to learn about voice agents with @carodivittorio ⤵️
Caroline di Vittorio
@carodivittorio
NYC 🗽 Join us on September 22 for a meetup on voice agents, where I'll be talking about practical ways to evaluate them!
RSVP here: luma.com/l27vmv25 -
LangChain
@LangChain
Just one week to go until the LangSmith Roadshow in Dallas!
Hear from @hwchase17, go hands-on with Deep Agents + LangSmith Engine, and meet the LangChain community!
RSVP today! events.langchain.com/register/e08624d8-a0f4-46eb-aa67-9e4… -
LangChain
@LangChain
We’re joining the @AWSstartups program as an Activate Exclusive Offers partner.
Now, eligible startups can now access up to $10,000 in LangSmith credits to help them build, test, evaluate, deploy, and operate agents on @awscloud.
Interested in getting started? Visit -
LangChain reposted
Caspar
@caspar_br
Connections offer a simple, elegant way to implement agent auth. We support agent and user (OBO) credentials in mda 0.7. Try it out, it’s magic langch.in/mda
LangChain
@LangChain
New in Managed Deep Agents 0.7.0+: Connections.
Agent-owned secrets for tools everyone shares, user-owned OAuth for anything that should carry a real person's identity.
Store your credentials in your LangSmith workspace instead of .env.
www.langchain.com/blog/connections-managed-credentials-an… -
LangChain
@LangChain
In 10 minutes, we set up Connections in Managed Deep Agents. Watch as we configure:
1️⃣ A shared @tavilyai key every user pulls from
2️⃣ A @Linear connection, OAuth via just a server URL
3️⃣ A GitHub OAuth app that files tickets under the user's own login, not a botThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
New in Managed Deep Agents 0.7.0+: Connections.
Agent-owned secrets for tools everyone shares, user-owned OAuth for anything that should carry a real person's identity.
Store your credentials in your LangSmith workspace instead of .env.
www.langchain.com/blog/connections-managed-credentials-an… -
LangChain
@LangChain
You can use the LangSmith CLI to pull your flagged traces, classify each one by issue type, and build an eval dataset with a split for each.
A 3 minute tutorial ⤵️This media couldn't be displayed. See it on the original post. -
LangChain reposted
Sydney Runkle
@sydneyrunkle
want to build a company brain like stripe?
they built it on deepagents, an open source harness. it’s super easy to customize, you can connect to any data source under the sun, and its model agnostic!
github.com/langchain-ai/deepagents
claire vo 🖤
@clairevo
The team at @stripe is setting the standard for internal AI platforms: minion coding agents, a custom prototyping rig, and now their company brain, Kai.
On today's episode of How I AI, Sharadh shows us how 1.5 engineers and 2 weeks got them a company brain, including:
- projects as governance
- skill routing + telemetry
- a skills platform that works for 10k teammates
Plus, he and I debate the merits of gentle parenting your AI (esp when your company is running evals.)
Full episode on YT: www.youtube.com/watch?v=AbZODZ_4VaMThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
.@WeberShandwick + @focused_dot_io are joining us at Interrupt NYC. interrupt.langchain.com/nyc
They'll share how a small delivery team built an agent factory on LangGraph, turning brand-tailored agent deployments from weeks of engineering into often <1hr of configuration.
Catch this talk + more @ Interrupt NYC. -
LangChain
@LangChain
Next stop, NYC!
-
LangChain
@LangChain
New in Managed Deep Agents 0.7: Connections
Caspar
@caspar_br
Agent auth is hard, but it shouldn't be!
Your agent needs to act as someone: sometimes that's one shared identity, sometimes it's the person asking. To get either right you'll need an OAuth app for each service, a token store, refresh logic, and a consent flow
Today we're fixing this with Managed Connections: define who your agent acts as in code, and the rest is handled for you:
𝚊𝚐𝚎𝚗𝚝 -> agent uses its own token
𝚞𝚜𝚎𝚛 -> agent uses the user's token, guiding them through the consent flow if they haven't connected yet
With a single argument you avoid the whole OAuth dance. Available now, in 𝚖𝚊𝚗𝚊𝚐𝚎𝚍-𝚍𝚎𝚎𝚙𝚊𝚐𝚎𝚗𝚝𝚜 0.7 -
LangChain reposted
Caspar
@caspar_br
Agent auth is hard, but it shouldn't be!
Your agent needs to act as someone: sometimes that's one shared identity, sometimes it's the person asking. To get either right you'll need an OAuth app for each service, a token store, refresh logic, and a consent flow
Today we're fixing this with Managed Connections: define who your agent acts as in code, and the rest is handled for you:
𝚊𝚐𝚎𝚗𝚝 -> agent uses its own token
𝚞𝚜𝚎𝚛 -> agent uses the user's token, guiding them through the consent flow if they haven't connected yet
With a single argument you avoid the whole OAuth dance. Available now, in 𝚖𝚊𝚗𝚊𝚐𝚎𝚍-𝚍𝚎𝚎𝚙𝚊𝚐𝚎𝚗𝚝𝚜 0.7 -
LangChain reposted
Harrison Chase
@hwchase17
harnesses should make context engineering easy
"forking" subagents is a useful context engineering trick that is now built into deepagents
LangChain OSS
@LangChain_OSS
-
LangChain
@LangChain
.@PodiumHQ's iteration loop runs through LangSmith end to end: user feedback becomes a trace, traces get reviewed, and recurring patterns get codified into LangSmith eval cases.
Walker Ward on why that beats playing whack-a-mole with bugs.
LangChain
@LangChain
Watch the full conversation: www.youtube.com/watch?v=J77ro1AJGa0
-
LangChain reposted
Sydney Runkle
@sydneyrunkle
subagents are great for parallelizing work and delegation to specialized agents, but it's hard to figure out exactly what context to pass to and return from subagents!
this guide details the two input context modes we now support
1. isolated -- subagent gets a completely new prompt
2. forked -- starts exactly where the main agent left off, with a copy of the old message history
as well as when you might want to use each!
LangChain OSS
@LangChain_OSS
-
LangChain reposted
Hunter Lovell
@huntlovell
Forking context is something you can do in a ton of different harnesses (claude, codex, now deepagents!) but it's something with very little guidance on when/how to use.
here's how we think about it!
LangChain OSS
@LangChain_OSS
-
LangChain
@LangChain
Adopting a coding agent for your org can mean a few less-than-optimal commitments at once:
🔒 Locking into a single vendor’s models
💲 Uncontrollable pricing plans + runaway token bills
🧠 Handing off your intelligence to a closed system
Dcode is open source, puts your team in control, and lets you own your intelligence.
Our latest LangChain Academy course shows you how.
LangChain
@LangChain
Deep Agents Code, or dcode, is an open source coding agent that works with any model + gives your team full control over usage.
In our latest LangChain Academy course, learn how to install + run dcode in a way designed for enterprise teams. academy.langchain.com/courses/dcode-essentialsThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Colin Francis
@colifran_
subagents now support forking in deep agents!
this gives your subagents the ability to start with all the relevant conversation history the supervisor agent has rather than starting blank.
check it out ⤵️
LangChain
@LangChain
New in Deep Agents: Subagents can now fork the supervisor's full conversation instead of starting from a blank context window.
Everything you need to know from @t_bengre and @chester_curme on the @LangChain_OSS team ⤵️
LangChain OSS
@LangChain_OSS
-
LangChain
@LangChain
New in Deep Agents: Subagents can now fork the supervisor's full conversation instead of starting from a blank context window.
Everything you need to know from @t_bengre and @chester_curme on the @LangChain_OSS team ⤵️
LangChain OSS
@LangChain_OSS
-
LangChain
@LangChain
78% resolution when a conversation is going well. 30% when it's not.
@GetCandidly's state model measures this gap in real time, and closes it.
One prompt insert mirroring the user's wording at the right turn can move a disengaging conversation back up.
LangChain
@LangChain
Full story: www.langchain.com/blog/how-candidly-built-state-aware-age…
-
LangChain
@LangChain
Deep Agents Code, or dcode, is an open source coding agent that works with any model + gives your team full control over usage.
In our latest LangChain Academy course, learn how to install + run dcode in a way designed for enterprise teams. academy.langchain.com/courses/dcode-essentialsThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
.@CopilotKit is sponsoring Interrupt NYC!
Stop by their booth during the event and go hands-on with the enterprise-ready frontend stack for agents.
LangChain
@LangChain
Learn more about Interrupt, The Agent Conference by LangChain and RSVP: interrupt.langchain.com
-
LangChain reposted
Jesse Palmer
@jesselpalmer
LangSmith has been working pretty well for me
-
LangChain
@LangChain
LangSmith for Startups: @raspberry__ai
✅ The agentic platform for fashion.
✅ Works alongside design teams on their boards and turns plain-English requests into finished renders, tech packs, and campaign imagery in minutes.
✅ Runs the full lifecycle of its LangGraph agent on LangSmith.
Join Raspberry AI to transform the fashion industry: jobs.ashbyhq.com/raspberry -
LangChain
@LangChain
You won't want to miss this -- @sydneyrunkle with a detailed breakdown of LangChain's MCP revamp and support for the new, stateless spec!
Sydney Runkle
@sydneyrunkle
ICYMI -- yesterday we released support for the new MCP protocol in @LangChain !! what do you need to know?
1. MCP support is now in the main langchain package! get started with `uv pip install 'langchain[mcp]'`
2. it's now built on top of FastMCP v4; FastMCP has ergonomic support for all things MCP (clients, servers, and apps). FastMCP handles things like protocol negotiation, tool name conflicts, connection lifecycle, transport specs, auth, and caching!
3. langchain now supports elicitation (MCP's human in the loop concept) via interrupts!
4. langchain now supports list_tools caching (so you don't have to re-fetch on every agent run; this was a common pain point for TTFT!)
5. the new protocol is stateless, which means the agent <-> server interactions are more efficient, and much more scale-able.
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
An amazing opportunity 👀
Ankush Gola
@ankush_gola11
We’re looking for someone to lead SmithDB at LangChain!
SmithDB is the database we’ve built from the ground up to power LangSmith. It’s optimized specifically for the access patterns that emerge when you’re storing and querying enormous volumes of agent traces.
It’s serving production traffic at significant scale with great performance, but it’s still early. The problems we are solving are genuinely fun and very hard: indexing, query execution, compaction, ingestion specifically for agent observability + making all of this fast and cost-efficient at massive scale.
We’re looking for an exceptional engineering leader to take ownership of this critical project. You’re likely a great fit if you’ve built database or distributed systems, led strong engineering teams through technically demanding projects, and have exceptional execution and project management skills.
Apply here or DM me directly: jobs.ashbyhq.com/langchain/d81c35c1-7243-45e4-bfe2-7fc874…
Learn more about SmithDB here: www.langchain.com/blog/introducing-smithdb
LangChain
@LangChain
PSA. We're hiring. www.langchain.com/careers
-
LangChain
@LangChain
Clay's agents are running longer and taking more steps.
LangSmith's threads feature is how they keep every one traceable.
LangChain
@LangChain
Watch the full conversation: www.youtube.com/watch?v=cx6_tb6HCeY
-
LangChain
@LangChain
New from the LangSmith Signal.
Over the last 2 weeks, we looked at which models teams reach for, and which ones are doing the work.
✅ Reach: gpt-4o-mini was used by 13% of orgs
✅ Work: gpt-4.1-mini sat at 7% of LLM calls
💡 DeepSeek V4 Flash: The only open-weight model to make either list, 2nd in reach at 9%.
LangChain
@LangChain
📊 We analyzed this information from LangSmith Observability data across billions of agent runs, and we're just getting started.
Stay tuned for more LangSmith Signals as we share how devs are building agents, by the numbers. www.langchain.com/langsmith-platform -
LangChain
@LangChain
.@RogoAI CEO & Co-Founder @GabeStengel is a headliner at Interrupt New York, The Agent Conference by LangChain.
See the agenda and get your tickets: interrupt.langchain.com -
LangChain
@LangChain
🎓 Get started with LangSmith Essentials: academy.langchain.com/courses/quickstart-langsmith-essent…
LangChain
@LangChain
The best orgs have figured out how to ship agents repeatedly, safely, and systematically.
They’ve established a continuous agent development lifecycle:
1️⃣ Build
2️⃣ Test
3️⃣ Deploy
4️⃣ Monitor
...So they can learn from real usage, and iterate quickly
With our free LangChain Academy course, LangSmith Essentials, you’ll explore the entire agent development lifecycle in less than 60 minutes.
Check it out: academy.langchain.com/courses/quickstart-langsmith-essent…This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
We have several LangChain-led and community-led meetups coming up! RSVP and bring a friend!
🇸🇪 9/8 Stockholm
🇺🇸 9/9 Pittsburgh
🇺🇸 9/15 Chicago
🇵🇱 9/15 Warsaw
🇵🇱 9/22 Wrocław
🗽 9/22 NYC
🇫🇷 10/5 Paris (w/ @hwchase17)
🇳🇱 10/6 Amsterdam (w/ Harrison Chase)
🇩🇪 10/27 Munich
RSVP 👉
LangChain
@LangChain
For the Pittsburgh meetup, please register here 👉 www.eventbrite.com/e/steel-city-ai-innovators-monthly-mee…
-
LangChain
@LangChain
ICYMI
Sydney Runkle
@sydneyrunkle
we now support the new stateless MCP spec in @LangChain!
we've revamped our entire MCP integration. It now lives in the main langchain package, elicitation works via interrupts, and list endpoints are cacheable.
the new adapter is built on top of FastMCP! FastMCP ships with first class transports, auth, connection management, protocol negotiation, caching, callbacks, and more.
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Donald Gossen - d/aicc
@dongossen
@LangChain agents can now pay for what they need mid-task. Delegate a card once, set the caps, and the agent buys credits, tops itself up, and pays new providers it discovers along the way…every purchase traceable in LangSmith.
Humans set the guardrails. Machines move the money.
LangChain
@LangChain
See what we cooked up: www.langchain.com/blog/agents-that-pay-how-nevermined-emp…
Nevermined
@Nevermined_ai
1/10 Agentic commerce has a missing layer.
Giving an AI agent permission to pay is not the same as knowing whether that payment was worth making.
That’s why we’re partnering with @LangChain.
🧵 -
LangChain
@LangChain
See what we cooked up: www.langchain.com/blog/agents-that-pay-how-nevermined-emp…
Nevermined
@Nevermined_ai
1/10 Agentic commerce has a missing layer.
Giving an AI agent permission to pay is not the same as knowing whether that payment was worth making.
That’s why we’re partnering with @LangChain.
🧵 -
LangChain
@LangChain
New cookbook with @Nevermined_AI:
✅ Let your agents pay for what they need mid-task, no human required
✅ Delegate your card, and set limits
✅ Let your agent buy credits, top itself up, and pay for the tools it uses
✅ Every purchase, traceable in LangSmith
Learn more: -
LangChain
@LangChain
Today, we're making some exciting updates to MCP in LangChain, including support for the new stateless MCP spec!
@SydneyRunkle with everything you need to know ⤵️
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Sydney Runkle
@sydneyrunkle
you can now build deepagents on top of @MongoDB Atlas!!
this backend, built on top of an underlying object store, supports more robust search including MongoDB search, vector search, and hybrid search!
check out the blog to learn more!
MongoDB
@MongoDB
MongoDB now powers the virtual file system behind @LangChain's Deep Agents. 🚀
This new integration lets you implement Deep Agents' BackendProtocol against MongoDB Atlas instead of building your own storage layer. And agent code calling read_file, write, glob, and grep don’t have to change.
Your retrieval and storage run through one backend instead of a maze of stitched-together tools. Vector, full-text, and hybrid search (via $rankFusion) route through Atlas; file reads and writes forward directly to S3.
For builders of long-running or multi-agent workflows: Plans, intermediate outputs, and knowledge artifacts persist across sessions, deployments, and sub-agent handoffs, so file-based context survives between runs instead of resetting each time. -
LangChain
@LangChain
The best orgs have figured out how to ship agents repeatedly, safely, and systematically.
They’ve established a continuous agent development lifecycle:
1️⃣ Build
2️⃣ Test
3️⃣ Deploy
4️⃣ Monitor
...So they can learn from real usage, and iterate quickly
With our free LangChain Academy course, LangSmith Essentials, you’ll explore the entire agent development lifecycle in less than 60 minutes.
Check it out: academy.langchain.com/courses/quickstart-langsmith-essent…This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
.@HeggieConnor's rule at @unifygtm: if the agent runs on GPT, the judge grading it needs to run on a different model family.
Choose the same family and you get mode collapse, or what is essentially groupthink for agents.
LangChain
@LangChain
Watch or listen to the latest Max Agency on your favorite podcasting platform.
🎧 Apple:
podcasts.apple.com/us/podcast/how-unify-cut-its-ai-agent-…
🎧 Spotify: open.spotify.com/episode/6kWQouc2QmiHGk0vdiZEtd
⏯️ YouTube: youtu.be/6898VdRtKDE -
LangChain
@LangChain
.@FactoryAI is sponsoring Interrupt NYC!
Stop by their booth during the event and go hands-on with the autonomy stack for enterprise teams.
LangChain
@LangChain
Learn more about Interrupt, The Agent Conference by LangChain and RSVP: interrupt.langchain.com
-
LangChain
@LangChain
In 2 weeks, the LangSmith Roadshow comes to Dallas!
✅ @hwchase17 keynote on the state of agent development.
✅ Hands-on workshops on Deep Agents + LangSmith Engine
Spaces are limited. RSVP today! events.langchain.com/register/e08624d8-a0f4-46eb-aa67-9e4… -
LangChain
@LangChain
Happening today!
An evening of technical talks, meeting builders, and end-of-summer vibes with @PrimeIntellect + @baseten!
luma.com/cwn8mze6 -
LangChain reposted
Sydney Runkle
@sydneyrunkle
@LangChain has day 0 support for Gemini 3.8 flash!
Patrick Loeber
@patloeber
Introducing Gemini 3.8 Flash, bringing significant improvements in agentic + coding capabilities from 3.7 Flash, available at the same price.
-
LangChain reposted
Brace
@BraceSproul
OpenWiki works best when paired with great coding agents. @colifran_ has been making it trivial to add to your favorite agent
Colin Francis
@colifran_
openwiki 🤝 cursor
we recently made openwiki more accessible than ever with coding agent integrations. our newest integration is cursor!
two commands and you're ready to go:
1⃣ npm install -g openwiki@latest
2⃣ openwiki integrations install <claude|codex|opencode|cursor>
new integrations are incredibly easy to add. let me know what you want to see next!
check it out ⤵️This media couldn't be displayed. See it on the original post. -
LangChain reposted
Colin Francis
@colifran_
openwiki 🤝 cursor
we recently made openwiki more accessible than ever with coding agent integrations. our newest integration is cursor!
two commands and you're ready to go:
1⃣ npm install -g openwiki@latest
2⃣ openwiki integrations install <claude|codex|opencode|cursor>
new integrations are incredibly easy to add. let me know what you want to see next!
check it out ⤵️This media couldn't be displayed. See it on the original post. -
LangChain reposted
Sydney Runkle
@sydneyrunkle
a core pillar of @LangChain oss has always been that we're model/provider agnostic
you can swap models at any time w/o having to worry about API differences :)
LangChain
@LangChain
“The model that is right for the job today is maybe not the model that’s right for the job next week.”
@sydneyrunkle on why we produce model and provider agnostic solutions. -
LangChain
@LangChain
OpenWiki now integrates with @cursor_ai!
Colin Francis
@colifran_
openwiki 🤝 cursor
we recently made openwiki more accessible than ever with coding agent integrations. our newest integration is cursor!
two commands and you're ready to go:
1⃣ npm install -g openwiki@latest
2⃣ openwiki integrations install <claude|codex|opencode|cursor>
new integrations are incredibly easy to add. let me know what you want to see next!
check it out ⤵️This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Before LangSmith deployments existed, @PodiumHQ hand-rolled and rebuilt their own agent runtimes.
Once LangSmith deployments existed, migrating was an easy call.
LangChain
@LangChain
Watch the full conversation: www.youtube.com/watch?v=J77ro1AJGa0
-
LangChain
@LangChain
“The model that is right for the job today is maybe not the model that’s right for the job next week.”
@sydneyrunkle on why we produce model and provider agnostic solutions. -
LangChain
@LangChain
We’re hosting a LangChain Amsterdam Meetup with @hwchase17 on October 6th!
Join us to learn how other companies have successfully leveraged the LangChain ecosystem, and hear directly from Harrison about the latest developments.
Space is limited, so RSVP today! -
LangChain reposted
Sydney Runkle
@sydneyrunkle
you can now build @LangChain agents w/ agentic video understanding!
google is consistently at the forefront of multimodal processing; this is the video analog of "agentic vision"
token usage + cost is also way down w/ this new approach!
docs: docs.langchain.com/oss/python/integrations/chat/google_ge…
Google
@Google
We’re introducing a new capability to our latest Gemini models: agentic video understanding.
This allows developers to process long-form video content with more accuracy, while using up to 88% less tokens.
See how it works 🧵This media couldn't be displayed. See it on the original post. -
LangChain reposted
Fireworks
@FireworksAI_HQ
Our partners @LangChain generate billions of tokens of agent traces a day, their richest signal on how real users react.
Judging every one with a frontier closed model like GPT-5.5 or Opus is too costly, so they fine-tuned a Qwen base model on Fireworks that matches it at up to 100x lower cost.
Build your own frontier: fireworks.ai/training -
LangChain
@LangChain
LangSmith Engine lets you see why the apple fell to the ground.
-
LangChain
@LangChain
Next stops on the LangSmith Roadshow:
📍 9/16: Dallas
📍 9/29: Boston
📍 10/1: Los Angeles
events.langchain.com/LangSmithRoadshow/
LangChain
@LangChain
The LangSmith Roadshow is heading to Boston!
✅ Learn about agent development + best practices in production.
✅ Take part in our workshops on agent harnesses with Deep Agents + improving your agent with LangSmith Engine
Spaces are limited!
events.langchain.com/LangSmithRoadshow/Boston/ -
LangChain
@LangChain
The LangSmith Roadshow is heading to Boston!
✅ Learn about agent development + best practices in production.
✅ Take part in our workshops on agent harnesses with Deep Agents + improving your agent with LangSmith Engine
Spaces are limited!
events.langchain.com/LangSmithRoadshow/Boston/ -
LangChain
@LangChain
Happening tomorrow!
Harrison Chase
@hwchase17
we're doing a meetup! "a summer evening on continual learning"
partnering with our friends at prime intellect and baseten
come join!
luma.com/cwn8mze6 -
LangChain
@LangChain
OpenWiki 0.5.0 is here!
@colifran_ with everything you need to know about our most durable and resumable release yet. 🧵
Colin Francis
@colifran_
🧵[1/3]
openwiki v0.5.0 is out 🎉
q: what's worse than long running jobs that fail?
a: long running jobs that fail and take all progress with it.
wiki generation takes time and tokens. openwiki respects your time and tokens.
the new architecture makes init and update durable and resumable:
➡️ completed pages are checkpointed as they finish
➡️ interrupted runs resume instead of restarting
➡️ partial progress survives ci failures
➡️ native openwiki and coding agent integrations all use the same resumable lifecycle -
LangChain
@LangChain
OpenAI's prompt cache makes a request 90% cheaper, but the cache key tops out around 15 requests per second. @HeggieConnor on how @unifygtm built its own routing around that limit, landing them close to a 95% cache hit rate.
LangChain
@LangChain
Watch or listen to the latest Max Agency on your favorite podcasting platform.
🎧 Apple:
podcasts.apple.com/us/podcast/how-unify-cut-its-ai-agent-…
🎧 Spotify: open.spotify.com/episode/6kWQouc2QmiHGk0vdiZEtd
⏯️ YouTube: youtu.be/6898VdRtKDE -
LangChain
@LangChain
.@focused_dot_io is sponsoring Interrupt NYC!
Stop by their booth during the event and meet the trusted LangChain partners building agents that integrate with your existing systems.
LangChain
@LangChain
Learn more about Interrupt, The Agent Conference by LangChain and RSVP: interrupt.langchain.com
-
LangChain reposted
Harrison Chase
@hwchase17
we're doing a meetup! "a summer evening on continual learning"
partnering with our friends at prime intellect and baseten
come join!
luma.com/cwn8mze6 -
LangChain reposted
Harrison Chase
@hwchase17
if only there was a company that helped you do both evals and harness engineering
elvis
@omarsar0
Next to evals, harness engineering is quickly becoming one of the most important skills for AI engineers to have today.
-
LangChain
@LangChain
🌉 Happening 9/2 in SF!
Join LangChain, @baseten, and @primeintellect for a dive into continual learning, what it means to own your own intelligence, and how teams are building systems that learn and improve over time.
luma.com/cwn8mze6 -
LangChain reposted
Sydney Runkle
@sydneyrunkle
we are rolling support for the new MCP spec in langchain and need feedback on our new API!
with this new pattern you connect langchain agents to multiple MCP servers, hook up oauth, use elicitation w/ langchain interrupts!
we're building on top of FastMCP which is the easiest way to get started building MCP clients and servers
github.com/langchain-ai/langchain/issues/40072 -
LangChain
@LangChain
Iconic
Caspar
@caspar_br
managed deep agents feels like magic. it handles the entire slack setup for you, and if you haven't picked an icon, it derives one from the agent's name. ~2M possible icons! inspired by dither-kit's DitherAvatar, thanks @grimcodes. langch.in/mda
-
LangChain reposted
Harrison Chase
@hwchase17
cost reconciliation at the trace level
For AI apps, "how much did we spend?" is less useful than "which workflow, tool call, prompt path, or retry pattern drove that spend?"
LangChain
@LangChain
Clay uses LangSmith to reconcile 99.5% of its AI spend. Head of AI @jeffbarg on why cost observability finally clicked once LangSmith was in place.
-
LangChain
@LangChain
Clay uses LangSmith to reconcile 99.5% of its AI spend. Head of AI @jeffbarg on why cost observability finally clicked once LangSmith was in place.
LangChain
@LangChain
Watch the full conversation: www.youtube.com/watch?v=cx6_tb6HCeY
-
LangChain
@LangChain
.@jpmorgan is speaking at Interrupt, the agent conference by LangChain.
See the full agenda and get your tickets: interrupt.langchain.com/nyc -
LangChain reposted
Harrison Chase
@hwchase17
Models labs will create great harnesses and ecosystems for their models but will block model access to harnesses of other labs
Only choice for a harness that works across models is one not associated with a lab
Long live LangChain
OpenAI
@OpenAI
We’re ending our partnership with Cursor following its acquisition by SpaceX. Under our proposal, Cursor’s direct access to our models would end on November 12.
We know that the people most affected by this decision are the developers who rely on OpenAI models in Cursor. We care about their experience in this transition and we’re ready to go above and beyond to support them.
openai.com/index/our-decision-on-cursor-following-its-acq… -
LangChain reposted
Harrison Chase
@hwchase17
Would love feedback on our support for the new mcp spec!
Sydney Runkle
@sydneyrunkle
we're adding support for the new MCP spec in @LangChain open source!
the new API is built on top of FastMCP, so you can take advantage of their excellent devx for building MCP servers and clients.
an early version is available in langchain==1.4.0a2; let us know what you think! -
LangChain
@LangChain
In 13 minutes, @jeffbarg, Vyshu Khota, and Soroush Khadem walk through how Clay scaled agent evals agents at 300M+ runs a month.
Topics covered:
✅ Their four quadrant eval framework
✅ Why closing the production-to-eval loop is the hardest part
✅ How a data lake and long context changed what agents can do with dataThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Two weeks before launch, @unifygtm's agent was burning so much compute that one message could instantly eat through a customer budget.
Co-Founder & CTO @HeggieConnor on how they cut costs by 90-95%
LangChain
@LangChain
Watch or listen to the latest Max Agency on your favorite podcasting platform.
🎧 Apple:
podcasts.apple.com/us/podcast/how-unify-cut-its-ai-agent-…
🎧 Spotify: open.spotify.com/episode/6kWQouc2QmiHGk0vdiZEtd
⏯️ YouTube: youtu.be/6898VdRtKDE -
LangChain reposted
LangChain OSS
@LangChain_OSS
new MCP spec support is coming soon in @LangChain (and deepagents)
try it out and let us know what you think!
Sydney Runkle
@sydneyrunkle
we're adding support for the new MCP spec in @LangChain open source!
the new API is built on top of FastMCP, so you can take advantage of their excellent devx for building MCP servers and clients.
an early version is available in langchain==1.4.0a2; let us know what you think! -
LangChain reposted
Sydney Runkle
@sydneyrunkle
we're adding support for the new MCP spec in @LangChain open source!
the new API is built on top of FastMCP, so you can take advantage of their excellent devx for building MCP servers and clients.
an early version is available in langchain==1.4.0a2; let us know what you think! -
LangChain
@LangChain
On September 17th, the LangChain Academy team is hosting a live workshop on Deep Agents.
RSVP today and get ready to learn how to build your own Deep Agent and see how planning, memory, and subagents let an agent run complex, multi-step tasks.
events.langchain.com/webinar/intro-to-deep-agents/ -
LangChain
@LangChain
Why @Box chose Deep Agents:
1️⃣ Complete model agnosticism
Customers can choose LLM providers, and Deep Agents allows this flexibility at the platform level.
2️⃣ Speed of iteration
Open agent harness = less time spent on building core agent infrastructure, more time solving enterprise-specific problems.
Inside the Box Agent ⤵️
www.langchain.com/blog/building-box-ai-how-an-enterprise-… -
LangChain
@LangChain
.@PodiumHQ's agents seemed broken.
LangSmith showed the real story: the agent was behaving rationally based on the context it had.
Principal Software Engineer Walker Ward on tracing agent reasoning end to end. -
LangChain
@LangChain
.@Airbnb is joining us at Interrupt NYC.
Pedro Rodriguez will share how Airbnb's Trust org went from prototype to a standardized production stack on LangChain and LangGraph, in a domain where being wrong is expensive.
Catch his talk + more @ Interrupt NYC, Sept 24. -
LangChain
@LangChain
LangSmith lets @MorningstarInc trace every agent decision in one place.
Senior Software Engineer Matt Trivett on what debugging looked like before and after LangSmith.This media couldn't be displayed. See it on the original post. -
LangChain reposted
LangChain JS
@LangChain_JS
🧑🍳
Christian Bromann
@bromann
Starting today, @LangChain’s Managed Deep Agents can bake your agent’s environment at deploy time. 🧑💻
No clone. No install. No repeated setup.
Drop in setup.sh or a Dockerfile. Bake once. Every new thread starts ready. 👇 -
LangChain reposted
Nick Hollon
@nickhollon10
@colifran_ has been absolutely cooking on openwiki! pretty impressive to see all the improvements he has made in such a short time
Colin Francis
@colifran_
the claims runtime in openwiki v0.4.0 gives your wiki a persistent source of truth to help it forget and self-correct.
but how would a reader know:
1⃣ what sources a page was built from?
2⃣ if a page is fully verified?
3⃣ who generated the page in the first place?
to answer that we migrated openwiki to support OKF v0.2
the claims runtime powers page-level sources and verification and openwiki separately stamps who generated the page and when.
claims=content-level truth
okf=page-level trust
special thank you to @seo_dev2142 for kicking off okf v0.2 support in openwiki! -
LangChain
@LangChain
These days it feels like everything is “agentic”, but what does it actually mean to build an agent?
In this free LangChain Academy course, you’ll learn agent engineering from the ground up, covering foundational concepts like react loops, MCP servers, and human in the loop as you go.
Check it out: academy.langchain.com/courses/foundation-introduction-to-…This media couldn't be displayed. See it on the original post. -
LangChain reposted
Colin Francis
@colifran_
the claims runtime in openwiki v0.4.0 gives your wiki a persistent source of truth to help it forget and self-correct.
but how would a reader know:
1⃣ what sources a page was built from?
2⃣ if a page is fully verified?
3⃣ who generated the page in the first place?
to answer that we migrated openwiki to support OKF v0.2
the claims runtime powers page-level sources and verification and openwiki separately stamps who generated the page and when.
claims=content-level truth
okf=page-level trust
special thank you to @seo_dev2142 for kicking off okf v0.2 support in openwiki! -
LangChain
@LangChain
New in OpenWiki 0.4.0: OKF v0.2 support
Colin Francis
@colifran_
the claims runtime in openwiki v0.4.0 gives your wiki a persistent source of truth to help it forget and self-correct.
but how would a reader know:
1⃣ what sources a page was built from?
2⃣ if a page is fully verified?
3⃣ who generated the page in the first place?
to answer that we migrated openwiki to support OKF v0.2
the claims runtime powers page-level sources and verification and openwiki separately stamps who generated the page and when.
claims=content-level truth
okf=page-level trust
special thank you to @seo_dev2142 for kicking off okf v0.2 support in openwiki! -
LangChain
@LangChain
New in Managed Deep Agents
Christian Bromann
@bromann
Starting today, @LangChain’s Managed Deep Agents can bake your agent’s environment at deploy time. 🧑💻
No clone. No install. No repeated setup.
Drop in setup.sh or a Dockerfile. Bake once. Every new thread starts ready. 👇 -
LangChain
@LangChain
Build an agent with natural language.
Deploy to @SlackHQ in one click.
More on our launch partnership for Add to Slack ⤵️
Slack
@SlackHQ
Now you can build in your favorite AI tool and deploy straight to Slack in one click. ⚡️
Get started with these partners:
@hyperagentapp
@LangChain
@Lovable
@n8n_io
@NanoClaw_AI
@OpenAI
@runlayer
@getskydive
@Superhuman
@vercel -
LangChain
@LangChain
Harnesses explained in 30 seconds by @sydneyrunkle.
LangChain
@LangChain
Watch the full Deep Agents session featuring @sydneyrunkle and @jakebroekhuizen.
youtu.be/GbzEDgcuGJU -
LangChain
@LangChain
New in OpenWiki 0.4.0: Coding agent integrations.
More from @colifran_ ⤵️
Colin Francis
@colifran_
a question we were interested in answering for openwiki 0.4.0 was: how do we make openwiki available to more devs? our answer was simple: coding agent integrations!
you can now drive openwiki init and update using claude, codex, and opencode with almost no set up needed:
1⃣ npm install -g openwiki@latest
2⃣ openwiki integrations install claude|codex|opencode
3⃣ "use openwiki to create a wiki for this repo" or "use openwiki to update my wiki for this repo"
go try openwiki. it is easier than ever.
watch me drive openwiki update using codex 👇This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
In just three minutes, see how LangSmith Engine can help you automate agent improvement by:
🔎 Identifying issues
💡 Proposing fixes
🔨 Building dataset examples
👀 Monitoring for regressionsThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Viv
@Vtrivedy10
every day @nickhollon10 wakes up, looks at the data, riffs on what it could mean, and makes cool Harbor tasks 🐐
-
LangChain
@LangChain
Inside @Rippling’s eval pipeline:
✅ Offline evals: Pre-recorded mocks + fixtures that run locally on every commit without external dependencies.
✅ Post-merge integration evals (online): 300-400 queries against a full Rippling sandbox to validate system health before deployment.
✅ Deploy-blocking evals (online): ~10 critical scenarios against real systems that gate every deployment.
✅ Continuous evals (online): Scheduled runs against prod data, multiple times daily, monitoring live system health.
LangChain
@LangChain
For @Rippling, LangSmith makes pulling and analyzing all conversations at scale simple.
“The ability to pull and analyze all conversations at scale… LangSmith makes that possible. We have a bunch of automated analysis running on top of it.”
— Laks Srini, Product Owner
www.langchain.com/blog/how-rippling-went-ai-native-across… -
LangChain reposted
Harrison Chase
@hwchase17
How we think about evaluating wikis
Including whether wikis are actually even helpful! -
LangChain reposted
Mason Daugherty
@masondrxy
We built WikiBench!
-
LangChain reposted
Nick Hollon
@nickhollon10
the evals grind continues! this time on OpenWiki. we built WikiBench to answer two things:
1. how good is a given wiki?
2. does the wiki help at all?
for the second, we ran the same questions three ways: wiki only, source only, and both. read on to find out which won. -
LangChain
@LangChain
.@unifygtm Founder + CTO @HeggieConnor on why his best eval habits came from self-driving.
LangChain
@LangChain
Watch or listen to the latest Max Agency on your favorite podcasting platform.
🎧 Apple:
podcasts.apple.com/us/podcast/how-unify-cut-its-ai-agent-…
🎧 Spotify: open.spotify.com/episode/6kWQouc2QmiHGk0vdiZEtd
⏯️ YouTube: youtu.be/6898VdRtKDE -
LangChain
@LangChain
Interrupt NYC is less than a month away.
This one-day event is made for AI leaders and engineers building and deploying agents in production.
Info + RSVP: interrupt.langchain.com/nyc -
LangChain reposted
Viv
@Vtrivedy10
word on the street is that Ben Franklin used LangSmith to mine his traces and turn them into environments to hill-climb agents for science
LangChain
@LangChain
LangSmith Engine is your lightning rod. Catch issues when they strike.
-
LangChain
@LangChain
LangSmith Engine is your lightning rod. Catch issues when they strike.
-
LangChain
@LangChain
Join us next week for The Learning Loop in SF on September 2nd: luma.com/cwn8mze6
Hear from:
✅ @jakebroekhuizen- LangChain
✅ @willcb- @PrimeIntellect
✅ @oneill_c- @baseten
We’ll dive into continual learning, what it means to own your own intelligence, and how teams are building systems that learn and improve over time.
After the talks, we’ll head to the patio where you can meet the LangSmith Engine team, enjoy food and drinks, slot car racing, an AI photo booth, and more.
Spots are limited, so be sure to register! -
LangChain
@LangChain
Accelerate every step of the agent development lifecycle:
🔎 Detect production issues
💡 Identify root causes
🔧 Propose fixes to prompts and code
🔄 Monitor for recurrences
LangChain
@LangChain
LangSmith Engine now offers >2x performance on key internal benchmarks.
Engine has already helped identify tens of thousands of issues in our customers’ agents. Now it offers:
✅ More accurate issue detection, clustering, and remediation
✅ Support for SaaS and self-hosted deployments
✅ Reduced Analysis mode for cost-sensitive customers
✅ Integrations with @slackhq and @Linear
✅ Automatic closing of stale issues
Learn more → www.langchain.com/blog/new-in-langsmith-engine-2x-better-… -
LangChain
@LangChain
LangSmith Engine now offers >2x performance on key internal benchmarks.
Engine has already helped identify tens of thousands of issues in our customers’ agents. Now it offers:
✅ More accurate issue detection, clustering, and remediation
✅ Support for SaaS and self-hosted deployments
✅ Reduced Analysis mode for cost-sensitive customers
✅ Integrations with @slackhq and @Linear
✅ Automatic closing of stale issues
Learn more → www.langchain.com/blog/new-in-langsmith-engine-2x-better-… -
LangChain reposted
Nick Hollon
@nickhollon10
@Vtrivedy10 and I have spent a lot of time thinking about how to efficiently build evals and environments. our eval-engineering skill detailed here helps spin the environment generation loop very quickly!
-
LangChain reposted
Harrison Chase
@hwchase17
its relatively easy to generate a wiki the first time, but how do you update them reliably?
openwiki v0.4.0 improves in this dimension - it "forgets" better
Colin Francis
@colifran_
openwiki v0.4.0 just went live 🎉
the focus? forgetting, accessibility, trust
1⃣ stale docs detection with forgetting and self-correction
2⃣ drive init and update through claude, codex, and opencode
3⃣ okf v0.2 support for provenance and verification
🧵1/4 -
LangChain reposted
Colin Francis
@colifran_
openwiki v0.4.0 just went live 🎉
the focus? forgetting, accessibility, trust
1⃣ stale docs detection with forgetting and self-correction
2⃣ drive init and update through claude, codex, and opencode
3⃣ okf v0.2 support for provenance and verification
🧵1/4 -
LangChain
@LangChain
.@colifran_ on when knowledge goes stale, and how OpenWiki can now catch this before you do.
Colin Francis
@colifran_
-
LangChain reposted
Viv
@Vtrivedy10
like everything else, it's a skill + many references, give it a try, we're using it for env/eval generation for our internal agents
www.skills.sh/langchain-ai/langchain-skills/eval-engineering -
LangChain reposted
Harrison Chase
@hwchase17
trying to think about how to most easily create evals
we launched a skill to help with this iterative process -
LangChain reposted
Viv
@Vtrivedy10
this is a practical guide (+ updated skill) on how we use real world data like traces + human feedback to make synthetic environments + evals so we can measure, harness engineer, and post-train our agents
a few main components:
- World Knowledge gathering of the systems our agent will interact with
- Using Specs to coordinate what will be built —> World Spec & Task Spec
- Human-Agent collaboration to iteratively edit a Spec and infuse human feedback
- An execution pipeline that takes an agreed on Task Spec and generates an environment + task
- Fixing design flaws in the task and environment by actually running agents in the environment & mining the verifier scoring + traces
@harrison_chase, @nick_hollon and I spent a ton of time going back and forth on design decisions and the overall flow
- how much a coding agent should prompt for human feedback
- what specs do we need and what do they contain
- how to create and update world knowledge
this new flow is packaged in our updated eval-engineering skill.
we want to help every team own their intelligence which means owning the pipeline to convert their valuable data into Tasks+Environments that improve their agent over time
check out the skill and reach out if you’re thinking about this! www.skills.sh/langchain-ai/langchain-skills/eval-engineering -
LangChain
@LangChain
.@Vtrivedy10 + @nickhollon10 on why you should separate deciding what a task should look like from building it, and how to let a coding agent turn what it learns into a reusable "world spec" for every future task.
A closer look ⤵️ -
LangChain
@LangChain
What does it take to move customer experience agents from an initial use case into a production system that improves over time?
Our latest guide brings together lessons from @Lyft, Fastweb + Vodafone, and @LATAMAirlines.
Download our guide to explore the full architectures, results, and lessons from each team: www.langchain.com/resources/customer-experience-cx-agents… -
LangChain
@LangChain
"LangSmith gives us the ability to monitor all of our agents, understand what's working, what's not, what tool calls have failed, what PR just broke the pipeline, and what features our users are adopting."
An inside look at how @Toyota went from shipping one production agent every 6 months to shipping 50+ using Deep Agents and LangGraph. www.langchain.com/blog/how-toyota-north-america-put-enter… -
LangChain reposted
Harrison Chase
@hwchase17
A hard part of agents is connections - seems easy but hard to get right/seamless
Managed deepagents tackles it for you
More to come!
Christian Bromann
@bromann
Starting today, a @LangChain's Managed Deep Agent deploy provisions your Slack app for you 🤯
No manifest. No OAuth redirects. No bot tokens to copy around.
One command, and your agent says 👋 in Slack. -
LangChain reposted
Alex Olsen
@its_ao
If you have had to set up a slack app, you know how nice it will be not to have to do that anymore
Christian Bromann
@bromann
Starting today, a @LangChain's Managed Deep Agent deploy provisions your Slack app for you 🤯
No manifest. No OAuth redirects. No bot tokens to copy around.
One command, and your agent says 👋 in Slack. -
LangChain
@LangChain
Join us for our Building Agents with Agents Roadshow 🏎️ in Sunnyvale next Thursday, August 27th.
RSVP: events.langchain.com/LangSmithRoadshow/SiliconValley/
- Keynote from @hwchase17 on the state of agent development and trending architectures
- Hear from Mani Najmabadi, Director of Machine Learning at Expedia on how they’re building agents
- Two conceptual deep dives on how to build with agent harnesses (Deep Agents) and how to fix your agents with agents (LangSmith Engine)
- Hands-on workshop on how to improve your agent leveraging a continuous improvement loop
- Meet other frontier agent builders. This event gets packed across our different stops around the country! 🌎️ -
LangChain
@LangChain
👀
Caspar
@caspar_br
just released managed deep agents 0.6.0. we now automate the ugly parts of deploying your agent to slack. it's like terraform for agents
-
LangChain reposted
LangChain JS
@LangChain_JS
You deploy, we provision! 🤝
Christian Bromann
@bromann
Starting today, a @LangChain's Managed Deep Agent deploy provisions your Slack app for you 🤯
No manifest. No OAuth redirects. No bot tokens to copy around.
One command, and your agent says 👋 in Slack. -
LangChain
@LangChain
👋
Christian Bromann
@bromann
Starting today, a @LangChain's Managed Deep Agent deploy provisions your Slack app for you 🤯
No manifest. No OAuth redirects. No bot tokens to copy around.
One command, and your agent says 👋 in Slack. -
LangChain
@LangChain
Help us make dcode even better!
More from @its_ao ⤵️
Alex Olsen
@its_ao
dcode users: let's chat!
We're looking for users who want to help make dcode the best coding agent in software engineering
We'll talk planned features, feedback (the good and the bad) and see about making dcode the best possible agent for you
DMs are open! -
LangChain
@LangChain
"The challenge of accuracy, the challenge of quality of answers, the challenge of guardrailing, are very real.
When you deploy a solution at scale, you need tooling like LangSmith.”
@SchneiderElec's story www.langchain.com/blog/how-schneider-electric-built-their… -
LangChain
@LangChain
One month to go until Interrupt NYC!
Save your spot at interrupt.langchain.com/nyc -
LangChain
@LangChain
At our recent SF meetup, @jacobtpl from @cognition walked through DeepWiki, the documentation tool behind Devin.
Watch his full session ⤵️This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Now you can learn LangChain using your favorite coding agent with LangChain Academy Tutors.
Tutors use skills to teach concepts, ask questions, run quizzes, and give feedback.
You can also customize the teaching style to match how you learn.
LangChain Academy: academy.langchain.com
LCA Tutors Repo: github.com/langchain-ai/lca-tutorsThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
.@Jimdo is joining us at Interrupt London.
VP of AI Jan Schlie will share how Jimdo's AI Companion for small businesses uses LangSmith to trace real product usage and turn those insights into safer, more relevant decisions.
Catch this + more: interrupt.langchain.com/london -
LangChain
@LangChain
It's super easy to build browser agents with stagehand from @browserbase and deepagents from @LangChain_OSS!
Browser agents can click, scroll, screenshot, extract structured data, and more. For example, @sydneyrunkle just built one to navigate maps!
docs.langchain.com/oss/python/integrations/tools/stagehand
Sydney Runkle
@sydneyrunkle
last night i spent some time building a browser agent with stagehand from @browserbase and deepagents from @LangChain
i wanted to build something that required pretty complex web navigation to prove the value of agentic browser use!
my agent plays a game called "map tap" where you are assigned a city around the globe, and have to drag the map, zoom in, and click as close to the given city as possible. you get points based on how close you are.
1. one shotted the agent w/ browserbase + langchain docs, it scored ~300/1000 points
2. asked my coding agent to review the trace and make it faster (smaller model) and more accurate (higher score)
3. came back 10 minutes later to see this in my trace:
here's a guide on how to build your own browser agent!
docs.langchain.com/oss/python/integrations/tools/stagehand -
LangChain
@LangChain
LangSmith for Startups Spotlight: Vector Legal
Vector Legal is the premier AI-native law firm for startups, venture capital, and private equity.
VectorOS, the one stop shop legal operating system, powers day-to-day legal work like formation, trademarks, contract management, company records, and e-signatures, while expert attorneys are on call when the stakes go up: fundraising, complex contract review, equity management, and bet-the-company matters.
Trusted by companies including Corgi, Browser Use, and Sela AI, Vector Legal recently raised a seed round led by Base10 Partners with participation from Y Combinator.
They use LangSmith for agent deployment, tracing, and fleet management across the AI agents powering their legal workflows.
Building something big, or backing those who are? Vector Legal would love to help: www.vectorlegal.com -
LangChain reposted
LangChain OSS
@LangChain_OSS
What's new in LangChain? 🚀
🧯 Standard exceptions: chat models now raise standard exception types across providers, so you can distinguish retryable errors (timeouts, rate limits, etc.) in a consistent way. Fully backward compatible.
🧩 Agent middleware: custom token_counter in ContextEditingMiddleware, plus non-retryable exceptions now skip retries in ModelRetryMiddleware (github.com/syyy44, github.com/Yigtwxx)
🔥 Fireworks document reranking: a new reranker integration (github.com/11adyy)
📊 Token & usage accounting: bug fixes for xAI and DeepSeek (github.com/aryansk, github.com/nazsats)
⚡ Portability & perf: lazy transformers import, and tighter grep scope for Anthropic (github.com/jtoman, github.com/Haaaarry)
This is your harness! Thanks to everyone who made it better. 🙏 -
LangChain
@LangChain
We are excited to be hosting Alex Atallah, Co-Founder & CEO of OpenRouter, at Interrupt NYC next month on September 24th!
See the agenda and get your tickets: interrupt.langchain.com/nyc -
LangChain
@LangChain
Join us for our Building Agents with Agents Roadshow 🏎️ in Sunnyvale next Thursday, August 27th.
RSVP: events.langchain.com/LangSmithRoadshow/SiliconValley/
- Keynote from @hwchase17 on the state of agent development and trending architectures
- Hear from Mani Najmabadi, Director of Machine Learning at Expedia on how they’re building agents
- TWO hands-on workshops: learning how to build with agent harnesses (Deep Agents) and improving your agent with agents (LangSmith Engine)
- Meet other frontier agent builders. This event gets packed across our different stops around the country! 🌎 -
LangChain
@LangChain
Preview Builds are now in Public Beta for LangSmith Deployment.
Test agent changes before merging by automatically spinning up temporary, production-like deployments from a PR branch.
Inspect traces, share feedback, and merge once the agent behaves the way you expect.
Preview Builds are short-lived and will automatically tear down after you merge changes.
Read the blog: www.langchain.com/blog/langsmith-preview-builds-test-agen…This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Join LangChain and Fireworks August 25th for a hands-on workshop where you'll learn to train a model yourself, followed by a rooftop happy hour.
Evals are the foundation for all system improvements, but building agentic systems is challenging. Human-in-the-loop evaluations are costly and time intensive, while LLM as a judge systems can be unreliable and subject to bias. In this workshop, you'll learn to build customized eval models that you can leverage in LangSmith to optimize your agent performance.
RSVP here: luma.com/phzy304n -
LangChain
@LangChain
Production conversations can reveal customer needs the original agent was not designed to handle.
@LATAMAirlines found that 13% of messages to its Concierge agent were being classified as out of scope. After reviewing production traces, the team discovered that 95% were legitimate passenger needs.
Adding a customer-care specialist reduced the out-of-scope rate from 13% to 1% and improved the return rate by 6%.
Learn how LATAM Airlines and other CX teams are using production feedback to improve agent performance: www.langchain.com/resources/customer-experience-cx-agents… -
LangChain reposted
Harrison Chase
@hwchase17
lots of observability & evals platforms
not a lot of platforms that help close the loop and have an agent that suggests fixes, adds evals itself, etc -
LangChain reposted
Sydney Runkle
@sydneyrunkle
come join one of the @lennysan 100 -- @LangChain !!
specifically, come work w/ me in person in SF and help devs build agents on open source
looking for someone who loves a steep learning curve and has strong opinions
www.langchain.com/careers?ashby_jid=74e5f9f4-e44a-4594-ba…
Lenny Rachitsky
@lennysan
3. The Lenny 100
To help you prioritize your search, I’ve curated a list of the top 100 companies I’d most want to join if I were looking for a job. These are companies that IMHO have the highest talent density, are tackling the most ambitious problems, have the biggest upside potential—and are actively hiring. I’ll update this list regularly: www.lennysjobs.com/lenny100 -
LangChain
@LangChain
We are excited to be featured as one of Lenny Rachitsky's top 100 companies Lenny would most want to join (if he was looking for a job today) -- companies with the highest talent density, tackling the most ambitious problems, and with the biggest upside.
We're hiring ➡️
Lenny Rachitsky
@lennysan
3. The Lenny 100
To help you prioritize your search, I’ve curated a list of the top 100 companies I’d most want to join if I were looking for a job. These are companies that IMHO have the highest talent density, are tackling the most ambitious problems, have the biggest upside potential—and are actively hiring. I’ll update this list regularly: www.lennysjobs.com/lenny100 -
LangChain reposted
Harrison Chase
@hwchase17
🎓New YouTube playlist: Managed Deep Agents
Gives an overview of Managed Deep Agents, and then each video dives deep into core concepts.
Launching with six videos!
1⃣ Intro: youtu.be/xdrB53bgpp0
2⃣ Conceptual Overview: youtu.be/TUJmfeGTr1Q
3⃣ Quickstart: youtu.be/L54dR9qMKzc
4⃣ Instructions and Context Hub: youtu.be/8HIoZV7qlwA
5⃣ Skills: youtu.be/Lhru0yMI2as
6⃣ Tools: youtu.be/piBwvWgHqkY
Playlist link: www.youtube.com/playlist?list=PLKYx7JmJ_IcA -
LangChain
@LangChain
RT @sydneyrunkle: you can now make your deepagent a browser agent w/ stagehand by @browserbase!
connect tools so your agent can click, typ… -
LangChain
@LangChain
.@Toyota builds its manufacturing agents on LangChain, and uses LangSmith to spot failed tool calls, broken pipelines, and frustration signals before they become bigger problems.
An inside look at how the two work together to keep production moving.This media couldn't be displayed. See it on the original post. -
LangChain reposted
Harrison Chase
@hwchase17
top tier webinar tmrw!
@Vtrivedy10 (@LangChain) @willcb (@PrimeIntellect) @AEllisBloor (@baseten) and I will will be jamming on how to automate more parts of the agent improvement loop (eval and environment engineering in particular)
Come join us: events.langchain.com/webinar/Towards-Automating-Eval-and-… -
LangChain
@LangChain
The Learning Loop in SF on September 2nd: luma.com/cwn8mze6
Hear from:
- Jake Broekhuizen, Labs Engineer, LangChain
- Will Brown, Head of Applied Research, @PrimeIntellect
- Charlie O'Neill, Co-Head of Model Training, @baseten
for an evening exploring continual learning, what it means to own your own intelligence, and how to build systems that learn and improve over time.
After the talks, we’ll head to the patio where you can meet the LangSmith Engine team, enjoy food, drinks, slot car racing, an AI photo booth, and more. Spots are limited. -
LangChain
@LangChain
What does it look like to automate eval and environment engineering for agents?
Join Harrison Chase, Vivek Trivedy, Will Brown, and Aaron Ellis-Bloor for Towards Automating Eval & Environment Engineering. We’ll explore how coding agents can use repository context and production traces to build evals, test them in reproducible environments, and keep developers in the loop.
Register: events.langchain.com/webinar/Towards-Automating-Eval-and-… -
LangChain
@LangChain
.@Modal CEO and Co-Founder @bernhardsson is a headliner at Interrupt NYC!
Join Erik and other industry leaders at our upcoming Agent Conference on September 24th. interrupt.langchain.com/nyc -
LangChain reposted
Sydney Runkle
@sydneyrunkle
we're hiring open source devs @LangChain
looking for people who are building at the frontier of agents and are excited to quickly develop ownership
apply: www.langchain.com/careers?ashby_jid=74e5f9f4-e44a-4594-ba… -
LangChain
@LangChain
Eval engineering livestream with Prime Intellect and Baseten ⬇️
Viv
@Vtrivedy10
come hang with a fun squad on how to turn your valuable data into environments to continuously improve models
covering fun stuff like:
- using real world trace data to create realistic models of the world
- how to think about environments & evals as training data for agents
- what does training actually look like? when should you do it
- how much do humans need to be in the loop at different parts of this cycle and why -
LangChain
@LangChain
An agent that fixes your agent. It’s genius.
-
LangChain
@LangChain
Join LangChain, @awscloud, + @MongoDB at the AWS Agentic AI Partner Showcase Extended in San Francisco to see how production-ready agents are built, from code to deployment.
Hear from technical experts, watch live demos, and connect with fellow builders. www.aicamp.ai/event/eventdetails/W2026082717 -
LangChain
@LangChain
Tuned Evaluators automatically add quality feedback to production traces.
They come with the tuned model, prompt, and managed infrastructure already set up, so you can start finding agent behavior that needs attention in a few clicks.This media couldn't be displayed. See it on the original post.
LangChain
@LangChain
Introducing LangSmith Tuned Evaluators
They automatically score agent behavior in production, starting with Perceived Error.
Perceived Error is one of the clearest signals that your agent is giving users a helpful experience.
In our benchmark, our specialized model outperformed every frontier model we tested and reduced evaluation cost by 82%.
Learn more: www.langchain.com/blog/introducing-langsmith-tuned-evalua… -
LangChain reposted
Jake Broekhuizen
@jakebroekhuizen
Perceived Error is one of the clearest signals that your agent is giving users a great experience.
We tuned the model and prompt around signals that tend to carry across domains, like a user correcting or rejecting an answer, repeating a request because something was missed, or leaving the conversation with an unresolved problem. Because these signals describe how the interaction unfolded rather than whether a domain-specific answer is objectively correct, we think Perceived Error maps really well across many types of conversational agents. Give it a try!
LangChain
@LangChain
Introducing LangSmith Tuned Evaluators
They automatically score agent behavior in production, starting with Perceived Error.
Perceived Error is one of the clearest signals that your agent is giving users a helpful experience.
In our benchmark, our specialized model outperformed every frontier model we tested and reduced evaluation cost by 82%.
Learn more: www.langchain.com/blog/introducing-langsmith-tuned-evalua… -
LangChain reposted
Harrison Chase
@hwchase17
🚀Today we launched LangSmith Tuned Evaluators, starting with Perceived Error.
Tuned Evaluators run on production traces to catch undesirable agent behavior and attach feedback that you can use in your agent improvement processes.
In our benchmark, our tuned model beat frontier models at 82% lower cost.
www.langchain.com/blog/introducing-langsmith-tuned-evalua… -
LangChain
@LangChain
Learn how @unifygtm cut 90-95% of model costs two weeks before launch on last week’s Max Agency episode
⏯️ YouTube: youtu.be/6898VdRtKDE
🎧 Apple: podcasts.apple.com/nz/podcast/how-unify-cut-its-ai-agent-…
🎧 Spotify: open.spotify.com/episode/6kWQouc2QmiHGk0vdiZEtdThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Introducing LangSmith Tuned Evaluators
They automatically score agent behavior in production, starting with Perceived Error.
Perceived Error is one of the clearest signals that your agent is giving users a helpful experience.
In our benchmark, our specialized model outperformed every frontier model we tested and reduced evaluation cost by 82%.
Learn more: www.langchain.com/blog/introducing-langsmith-tuned-evalua… -
LangChain
@LangChain
As model capabilities converge, more of the differentiation may come from the system around the model: the harness, tools, environments, evals, and feedback loops used to improve it.
Join @hwchase17, co-founder and CEO of LangChain, @Vtrivedy10 Applied Researcher at LangChain, @willcb , Applied Researcher at Prime Intellect, and Aaron Ellis-Bloor Applied Researcher at @baseten for a live panel session on Automating Eval & Environment Engineering.
We’ll explore the improvement loop and what open models and model-harness co-design mean for companies that want to own their intelligence.
Register now: events.langchain.com/webinar/Towards-Automating-Eval-and-… -
LangChain
@LangChain
In this practical webinar, we’ll show how to evaluate voice agents across execution, outcomes, and experience.
See how to build a more complete evaluation approach for improving voice agents in production.
Save your spot: events.langchain.com/webinar/How-to-evaluate-voice-agents/ -
LangChain
@LangChain
.@mondaydotcom built their AI copilot, Sidekick, to do more than answer questions.
It can handle complex, iterative work like analyzing CSVs or generating a map on the fly. To do this, they give agents an isolated workspace with LangSmith Sandboxes.
How they built it and what they learned: www.langchain.com/blog/building-monday-com-sidekick-why-c… -
LangChain
@LangChain
Your agents can now pay for the APIs they use.
We worked with AWS on the AgentCore Payments middleware. When a tool hits a paid API that returns a 402, it checks the price against your session budget, signs the payment, and retries. The agent gets its data and never has to handle the paywall.
Every payment also lands in LangSmith next to the reasoning that caused it. It's not enough that the agent worked and paid, somebody is going to ask what it bought and why.
Link to blog - www.langchain.com/blog/langchain-agentcore-payments -
LangChain
@LangChain
Headed to Interrupt? Don’t miss our Deep Agents from Scratch workshop.
Come learn about agent harnesses and build your first Deep Agent.
📍 NYC: Sept 24
📍 London: Oct 13
🎟️ interrupt.langchain.com -
LangChain reposted
Caspar
@caspar_br
anatomy of a managed deep agent. langch.in/mda
-
LangChain reposted
LangChain OSS
@LangChain_OSS
What's new in LangChain? 🚀
🔌 Support for OpenAI's 3.0 SDK (using httpx2)
✨ Support for gemini-3.7-flash
Plus a wave of core reliability fixes from external contributors:
🛠️ Tool calling & structured output: clearer errors, respect for pydantic aliases when validating tool inputs (github.com/RinZ27, github.com/kinch-tech, github.com/charan16)
📊 Usage metadata fixes: token usage callback bugs, and cost metadata when streaming OpenRouter (github.com/Solaris-star, github.com/iroiro147)
🧱 Content & prompt hardening: guard malformed Anthropic content blocks, and preserve non-str/non-dict items in prompt templates for content blocks (github.com/Oxygen56, github.com/zerafachris)
Huge thanks to everyone who contributed! 🙏 -
LangChain
@LangChain
In 17 minutes, @BraceSproul runs through everything you need to know about OpenWiki, an open source self-maintaining wiki.
✅ The core ideas behind it
✅ How it works
✅ Why docs built for agents look nothing like docs for humans
✅ What we learned from building itThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Reining in LLM spend gets harder once your team has started using multiple agents calling multiple models.
Without the proper controls, costs can spike before you realize what happened.
In this walkthrough, learn how to set and enforce consistent cost controls using LangSmith LLM Gateway.
Watch the demo:This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Curious about what’s new across LangChain and LangSmith?
Join our upcoming Customer AMA for a general overview of our latest product launches, updates, and capabilities.
events.langchain.com/register/customer-ama-august/ -
LangChain reposted
Caspar
@caspar_br
you can build agents that prompt themselves by giving them a schedule. with Managed Deep Agents, that only takes a few lines of code
code snippet is from our shiny new MDA docs: langch.in/schedules -
LangChain reposted
Caspar
@caspar_br
made a YT tutorial on how to build a social media agent with Managed Deep Agents: youtu.be/OpFXXSsEIBo
it scans Hacker News and X every morning and Slacks me interesting content to post about
let me know what sort of examples you want to see next!! -
LangChain
@LangChain
Improving agents is a data mining problem ⬇️
Viv
@Vtrivedy10
the @aiDotEngineer World Fair always one of the best events every year to talk to builders at the frontier of Research, Agents, Evals, Systems, etc
A few weeks ago I gave a talk on
- Continually Improving Agents
- building Agents to understand data from other Agents
- & a walkthrough of some of our latest work on data agents & post-training experiments
some fun takes:
- Every Continual Learning company will be an Observability & Eval company (and vice versa)
- Environments & Evals are the currency of agent improvement. Agents are literally following the behaviors encoded in Evals. The best way to make good evals is mining Production data at scale
- A good recipe to own your intelligence is using a Harness Eng - PostTrain - Harness sandwich with open models
- Model-Harness-Task fit! There is no universal model or universal harness. You can always build a better agent system by optimizing the model and harness for a given task
if your team is looking to understand your data at scale, build environment/evals, or just improve your agents - reach out, hmu would love to work with you! 🚀
youtu.be/CvRngaQZQ3Y -
LangChain
@LangChain
Next stop on the LangSmith Roadshow: Silicon Valley.
✅ A half-day, hands-on look at LangSmith Engine
✅ Hear from @hwchase17, @Expedia, + more
✅ Get hands-on experience building + deploying
RSVP today: events.langchain.com/LangSmithRoadshow/SiliconValley/ -
LangChain
@LangChain
LangSmith is now available as a bring your own cloud deployment on @awscloud.
Keep agent traces and runtime data inside your AWS boundary, while LangChain manages provisioning, upgrades, scaling, and support.
BYOC gives teams data sovereignty with the speed of a managed service.
www.langchain.com/blog/langsmith-byoc-is-now-generally-av… -
LangChain
@LangChain
📊We’ve rebuilt Dashboards in LangSmith to make them more useful for investigation and reporting.
Put KPIs next to trends, compare metrics with different units, break traces down by model or user, add notes, and arrange the dashboard freely.
docs.langchain.com/langsmith/dashboards -
LangChain
@LangChain
.@LinkedIn is joining us at Interrupt NYC.
Tanvi Motwani will share the Agentic Eval Playbook behind LinkedIn's Hiring Assistant: turning agent traces into golden datasets, training LLM-as-a-Judge systems, and building offline/online monitoring for production agents.
Catch her talk + more @ Interrupt NYC, Sept 24. interrupt.langchain.com/nyc -
LangChain reposted
LangChain OSS
@LangChain_OSS
New langchain python release! 🚀 Highlights:
• Internal middleware model calls no longer leak into messages / event stream (github.com/devbyteai, reported by github.com/nick-youngblut)
• SummarizationMiddleware retries failed summaries (reported by github.com/Harshxth) + faster token counting for tool schemas
• Batch of langchain-anthropic bug fixes (github.com/Mohith26, github.com/Functionhx)
• OpenRouter preserves provider info in response metadata (github.com/broskell)
This release also includes cross-provider standardization of reasoning effort, including supported values per-model in model.profile: docs.langchain.com/oss/python/langchain/models
Huge thanks to our external contributors for helping to shape and improve our harness! -
LangChain
@LangChain
.@Toyota built their R&D GPT on Deep Agents to help employees search years of paint, seat, and manufacturing research with a single command.
Here's their story. -
LangChain
@LangChain
Do all of your agents need a frontier model?
We ran NVIDIA's open source router, Switchyard, through our Deep Agents eval suite comprising of 145 multi-step tasks - Tool use, retrieval, filesystem operations, long context.
The result: 93% of turns went to a 30B model. Only 7% needed Claude Opus 4.8.
Routing between the two cut total cost by ~70%, while retaining ~90% of Opus's accuracy on the same calls.
Try out the deepagents integration with NVIDIA's Switchyard today: docs.langchain.com/oss/python/integrations/middleware/nvidia
LangChain
@LangChain
Link to blog: www.langchain.com/blog/switchyard-agent-routing-benchmark
-
LangChain
@LangChain
Already building voice agents?
Join our upcoming webinar to learn how to evaluate them across execution, outcomes, and experience—from tool use and task completion to latency, interruptions, and conversational friction.
events.langchain.com/webinar/How-to-evaluate-voice-agents/ -
LangChain reposted
Harrison Chase
@hwchase17
💻How to build a web browsing agent
Stagehand v4 is an SDK to let agents browse the web from @browserbase
We created a tutorial to show how you could use Stagehand with Managed Deep Agents to create a production ready web browsing agent
youtu.be/O0hkpChFBkM -
LangChain
@LangChain
🛡️Data Protection Controls with LangSmith LLM Gateway
✅ Detect, redact, and replace PII and secrets in requests before they reach model providers
✅ Redact sensitive data from the traces Gateway records in LangSmith
Now available in the LLM Gateway Public Beta for Enterprise users. Request access:
www.langchain.com/langsmith-llm-gateway-data-protection-r… -
LangChain
@LangChain
On the latest Max Agency, @cognition president @russelljkaplan shared why devs stopped chasing the best model and started chasing efficiency.
⏯️ YouTube: www.youtube.com/watch?v=bBUotstDLdk
🎧 Apple: podcasts.apple.com/us/podcast/the-misaligned-incentives-b…
🎧 Spotify: open.spotify.com/episode/7JNB5YXA6NVgUDAhy8eA3V -
LangChain
@LangChain
Interrupt is the place to connect with agent builders.
Whether-in between sessions, at our workshops, or during our receptions, there isn’t a better place to meet the builders, engineers, and teams shaping what's next for agents.
Interested in attending? RSVP while spaces last: interrupt.langchain.com -
LangChain reposted
Sydney Runkle
@sydneyrunkle
the term “RLM” (recursive language model) got a lot of buzz this week, but this idea is not new!
@a1zhang wrote the og RLM paper 10 months ago! thats like 5 agent-years!
would highly recommend following him if you want to stay ahead of the game on harness trends -
LangChain reposted
Sydney Runkle
@sydneyrunkle
@LangChain deepagents supports superserve sandboxes!
superserve sandboxes support automatic checkpointing, near instant init/teardown (<50ms), and the runtime is open source!
they also have no session limits, which is great for long running or ambient tasks!
github.com/superserve-ai/langchain-superserve -
LangChain reposted
Sydney Runkle
@sydneyrunkle
"RLM harness" is a buzzy term right now, what does it mean?
RLM = recursive language model. In practice this means a supervisor model can break down a task into smaller pieces and recursively call itself on those smaller tasks.
Here's how you can use RLMs w/ deepagents! -
LangChain reposted
Sydney Runkle
@sydneyrunkle
MDA: you bring the business logic / knowledge, we bring the intelligent harness and managed runtime
Harrison Chase
@hwchase17
-
LangChain
@LangChain
Austin Berke, Lead AI Product Engineer at @Harmonic recently shared how they rebuilt Scout on Deep Agents, quadrupling retention along the way.
Watch the full fireside chat at youtu.be/pGdZBK___jM -
LangChain reposted
Harrison Chase
@hwchase17
managed deepagents is one of the launches ive been more excited about in a while
wrote a bit about the journey from early langchain to managed agents, and why i think they will be a step change in how easy it is actually run agents
Harrison Chase
@hwchase17
-
LangChain
@LangChain
Managed Deep Agents is now in public beta.
Go from prototype to production scale without managing the underlying infrastructure yourself.
Built for developers who want to increase product velocity with control over model choice and the agent development lifecycle.
LangChain
@LangChain
From your terminal, you can:
→ Scaffold a Deep Agent with primitives for channels, sandboxes, memory, and identity
→ Deploy it to LangSmith-managed infrastructure
→ Run evals to verify agent behavior with @harborframework -
LangChain
@LangChain
🔍Tracing Controls with the LangSmith platform
When a Langsmith LLM Gateway guardrail triggers, it shows up in your LangSmith traces. Enforce and track spend limits, rate limits, and sensitive data controls in one platform. Then, open the trace to see what happened and make improvements.
www.langchain.com/blog/langsmith-llm-gateway-runtime-cont… -
LangChain
@LangChain
New from The LangSmith Signal: Open-weight models are on the rise, and ready for production.
✅ 1 in 3 active teams ran an open weight model in July, up from 1 in 5 years ago.
✅ Teams running an open-weight model daily grew 9x YoY, while closed models usage grew 4.8x YoY.
LangChain
@LangChain
📊 We analyzed this information from LangSmith Observability data across billions of agent runs, and we're just getting started. Stay tuned for more LangSmith Signals as we share how devs are building agents, by the numbers.
www.langchain.com/langsmith-platform -
LangChain
@LangChain
"SWE-bench is totally saturated."
@russelljkaplan on why @cognition built its own coding benchmark.
⏯️ YouTube: www.youtube.com/watch?v=bBUotstDLdk
🎧 Apple: podcasts.apple.com/us/podcast/the-misaligned-incentives-b…
🎧 Spotify: open.spotify.com/episode/7JNB5YXA6NVgUDAhy8eA3V -
LangChain
@LangChain
.@mondaydotcom is returning to Interrupt!
Director of Engineering Dor Cohen will share how monday.com built a dual-pipeline LLM eval system, using LangSmith production traces to drive both online and offline evaluation.
Catch Dor's talk + more @ Interrupt London, Oct 13. interrupt.langchain.com/london -
LangChain
@LangChain
LangSmith for Startups Spotlight: @Backopsai
✅ An AI-native resolution layer for supply chain and complex operations.
✅ Their agents handle operational workflows end to end: filing claims, reading emails, and updating systems like a top ops manager rather than a fragile script.
✅ Uses LangSmith for end-to-end workflow tracing, monitoring, and experimentation tools for their multi-step agent runs, helping them run evaluators and track performance metrics to catch regressions, before customers ever notice.
BackOps is hiring. Check out their open roles at backops.ai/careers. -
LangChain reposted
Ankush Gola
@ankush_gola11
We're hosting an @ApacheDataFusio meetup at our new SF office!
RSVP here: luma.com/gk3s2vib?tk=Bx03Pk
LangChain
@LangChain
On August 13, join us for a San Francisco @ApacheDataFusio Meetup hosted by @ankush_gola11.
You'll learn about the trajectory of modern data infrastructure, real-world applications, performance integrations, and more.
RSVP at luma.com/gk3s2vib -
LangChain
@LangChain
On August 13, join us for a San Francisco @ApacheDataFusio Meetup hosted by @ankush_gola11.
You'll learn about the trajectory of modern data infrastructure, real-world applications, performance integrations, and more.
RSVP at luma.com/gk3s2vib -
LangChain
@LangChain
The most common question we get about our open source is when to reach for each of Deep Agents, LangChain, and LangGraph
@sydneyrunkle just wrote up a guide on what each package offers, and when to reach for each, with examples!
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
📦 Try Out Kimi K3 in LangSmith LLM Gateway
Run agents with Kimi 3 using just your LangSmith API key.
LangSmith now provides access to open models out-of-the-box, powered by @FireworksAI_HQ.
Get started today → docs.langchain.com/langsmith/llm-gateway-credits -
LangChain
@LangChain
LangChain Academy Course update: Intro to Deep Agents
✅ Modules on dynamic and async subagents
✅ Course tutor skill which turns your coding agent into a personal tutor
✅ Capstone projects + practice exercises
✅ Support for Typescript
✅ Updates in Deep Agents v0.7
💻 Start now: academy.langchain.com/courses/foundation-introduction-to-… -
LangChain
@LangChain
✅ Just added: ISO 27001:2022 certification
✔️ SOC 2
✔️ GDPR
✔️ HIPAA compliance -
LangChain
@LangChain
"A lot of the bottleneck in chip design, it's actually thermals."
@cognition president @russelljkaplan on the Tesla debate that shaped how he thinks about inference.
⏯️ YouTube: www.youtube.com/watch?v=bBUotstDLdk
🎧 Apple: podcasts.apple.com/us/podcast/the-misaligned-incentives-b…
🎧 Spotify: open.spotify.com/episode/7JNB5YXA6NVgUDAhy8eA3V -
LangChain
@LangChain
Our latest guide outlines how teams can build the right governance foundation across cost, control, and compliance, including where gateway infrastructure can help centralize policy enforcement, data protection, rate limits, fallbacks, and spend management.
Read more: -
LangChain
@LangChain
LangSmith traces let the @MadrigalPharma team debug agents like reading a brain scan.
One click deploys can replace months of custom orchestration.
A look at how they took their agent time to production from 12 weeks to 2.This media couldn't be displayed. See it on the original post. -
LangChain reposted
-
LangChain
@LangChain
OpenWiki now includes a web visualizer for humans to utilize the wikis it generates.
Read the wikis, and inspect how repo sections relate to one another via the built in graph visualizer.
OpenWiki is an open source agent CLI for writing and maintaining repo documentation:
Brace
@BraceSproul
OpenWiki now ships with a built in web visualizer!
Run `openwiki visualize` to automatically spin up a UI to read your docs, inspect relationships between files via the graph viewer, and more!
Try it out locally today: github.com/langchain-ai/openwiki -
LangChain reposted
Brace
@BraceSproul
OpenWiki now ships with a built in web visualizer!
Run `openwiki visualize` to automatically spin up a UI to read your docs, inspect relationships between files via the graph viewer, and more!
Try it out locally today: github.com/langchain-ai/openwiki -
LangChain reposted
Sydney Runkle
@sydneyrunkle
i tweeted yesterday that i think orgs are going to be built on harnesses! it was surprisingly popular, so sharing more on what i meant!
savvy companies build agents that are:
* built on a core harness connecting a model to relevant data (slack, notion, ramp, gmail, salesforce, the list goes on...)
* accessible anywhere you work (slack, notion, github, gmail, zoom)
* increasingly multi-player
* both proactive (on schedules) and reactive (to live events) -
LangChain
@LangChain
👤Customer-Specific Controls with LangSmith LLM Gateway
Set rate and spend limits for each customer and user under a single API key. Keep usage predictable, simplify customer management, and avoid surprise model bills.
www.langchain.com/blog/langsmith-llm-gateway-runtime-cont… -
LangChain
@LangChain
Prototyping an agent takes an afternoon. Getting it to production doesn't.
The bottleneck is usually engineering speed, how fast you can catch issues and ship fixes. LangSmith Engine does that automatically, monitoring production traces & deploying fixes on its own.
Take our LangChain Academy course: academy.langchain.com/courses/langsmith-engine -
LangChain
@LangChain
“Skills are fancy prompts, but they’re really easy to share, and they can help agents really narrow in and do a certain task well”
@sydneyrunkle on why Skills matter.
LangChain
@LangChain
Watch the full fireside chat with @sydneyrunkle and @jakebroekhuizen
youtu.be/GbzEDgcuGJU -
LangChain
@LangChain
.@Spring_Health is joining us at Interrupt NYC. interrupt.langchain.com/nyc
Sr. Director, Engineering Marissa Saunders will share how they use LangSmith to catch risk on both sides of a mental health conversation: crisis signals from the user, and unsafe or hallucinatory responses from the model.
Catch her talk + more @ Interrupt NYC, Sept 24. -
LangChain reposted
Viv
@Vtrivedy10
Synthetic Environment Generation with Human Feedback
Agents are poor 1-shot eval/environment generators because they're misaligned with human goals, successful environments come from strategically infusing human feedback in the Env/Eval gen process
In the eval-engineering skill, we do this by aligning on 3 files
- Harness.md
- Environment.md
- Task.md
the flow is:
1. install the skill in your agent of choice: npx skills add github.com/langchain-ai/langchain-skills --skill eval-engineering
2. point it to your repo + traces (if you have them)
3. it will map the entire agent surface and propose eval directions based on real production data if you have traces. you can give feedback & propose updates on what the agent should be good at and how your org thinks about measuring this
4. it will generate Harness, Env, Task md files to align. Questions like what tools to simulate, the type of database to create in the environment, the definition of success that the verifier should check
5. Generate the Env/Task and run it in a loop to see how a real agent tackles it. Fix any broken piece of the eval and get user feedback in each step. A lot of this can only be discovered by running a real agent
nothing is hidden, everything lives locally in your environment and at the end you get a set of human-aligned Harbor Tasks
we want to help every team own their data & evals. it is largely a set of skills with very good guidance, good open sourcing tooling like Harbor, and strong human-agent collaboration
try it out and tweak as you need! -
LangChain reposted
Sydney Runkle
@sydneyrunkle
ran a webinar on loop eng
best part of this is the questions at the end!
www.youtube.com/watch?v=jPPiZ22DY3g -
LangChain
@LangChain
Learn how teams can establish the right governance foundations, centralize critical controls, and scale agents without losing visibility into cost, data, or policy enforcement.
events.langchain.com/register/building-governed-agents/ -
LangChain
@LangChain
Our latest guide, Customer Experience Agents in Production, looks at how @Lyft, @fastwebvodafone, and @LATAMAirlines are building agents that resolve customer requests, assist frontline teams, and improve through production feedback.
Read the guide: www.langchain.com/resources/customer-experience-cx-agents… -
LangChain
@LangChain
.@carodivittorio put together a framework on how to evaluate voice agents:
1️⃣ Execution: Did the agent follow its instructions?
2️⃣ Outcome: Did the interaction achieve its goal?
3️⃣ Experience: Was the conversation a smooth experience for the caller?
An inside look: -
LangChain
@LangChain
OpenWiki's code init prompt just got a major upgrade!
It now generates higher quality wikis that cover more of your codebase, leading to better eval scores and more efficient runs.
Try it out locally today: github.com/langchain-ai/openwiki
Brace
@BraceSproul
I just rewrote OpenWiki's entire prompt in v0.3 to generate more detailed & accurate wikis. Eval results:
- 📈 28.57% success increase (35% to 45% @ n=2)
- 🪙 14% fewer tokens
- 🧰 26% fewer tool calls
per successful task
And as you can clearly see from the graphs, the wikis contain much more data now.
Try out the new and improved OpenWiki: github.com/langchain-ai/openwiki
npm install -g openwiki@0.3.0 -
LangChain reposted
Brace
@BraceSproul
I just rewrote OpenWiki's entire prompt in v0.3 to generate more detailed & accurate wikis. Eval results:
- 📈 28.57% success increase (35% to 45% @ n=2)
- 🪙 14% fewer tokens
- 🧰 26% fewer tool calls
per successful task
And as you can clearly see from the graphs, the wikis contain much more data now.
Try out the new and improved OpenWiki: github.com/langchain-ai/openwiki
npm install -g openwiki@0.3.0 -
LangChain
@LangChain
🔄 Model Fallbacks with LangSmith LLM Gateway
Don’t let model outages take your agents with them.
Define fallback rules across models and hosts and use them across every agent. When a provider goes down or hits a rate limit, calls are routed to another model. -
LangChain
@LangChain
On the Max Agency podcast, @cognition president @russelljkaplan shared how his team builds proactive agents.
⏯️ YouTube: www.youtube.com/watch?v=bBUotstDLdk
🎧 Apple: podcasts.apple.com/us/podcast/the-misaligned-incentives-b…
🎧 Spotify: open.spotify.com/episode/7JNB5YXA6NVgUDAhy8eA3V -
LangChain
@LangChain
Proud to deliver additional resilience capabilities to our open source frameworks: Deep Agents, LangGraph, and LangChain.
Our work will allow agents to retry interrupted work, follow a safe recovery path, resume from a saved state instead of starting over, and automatically fall back to alternative models when the primary model fails.
NVIDIA
@nvidia
Security is strongest when organizations collaborate openly and move quickly.
Now over 120 members strong, the Open Secure AI Alliance is sharing new open source security contributions, including new proposed SAFE guidelines, for turning confidential incident findings into stronger protection across the ecosystem. -
LangChain reposted
Sydney Runkle
@sydneyrunkle
we recently rolled out new docs on fault tolerance for agents
everything from handling transient errors to engaging a human in the loop when information is missing
here's a guide on how to handle errors in agentic systems:
docs.langchain.com/oss/python/deepagents/fault-tolerance -
LangChain reposted
Harrison Chase
@hwchase17
Internal AI platforms have the potential to transform how companies function
Read how Stripe built theirs
LangChain
@LangChain
"My @Stripe career is divided into before and after Kai."
Stripe built Kai, their Knowledge AI Platform, with Deep Agents in one week with just one engineer.
Here’s the full story: www.langchain.com/blog/how-stripe-built-their-knowledge-a… -
LangChain
@LangChain
"My @Stripe career is divided into before and after Kai."
Stripe built Kai, their Knowledge AI Platform, with Deep Agents in one week with just one engineer.
Here’s the full story: www.langchain.com/blog/how-stripe-built-their-knowledge-a… -
LangChain
@LangChain
BYOK with LangSmith LLM Gateway
Take advantage of all that LangSmith LLM Gateway has to offer (fallbacks, cost controls, rate limits) while bringing your own keys.
Leverage your existing contracts and avoid paying a markup for tokens.
www.langchain.com/blog/langsmith-llm-gateway-runtime-cont… -
LangChain
@LangChain
.@carodivittorio joined @googledevs to share how you can
✅ Build voice agents with Google ADK + Gemini Live
✅ Set up tracing in LangSmith
✅ Get full visibility into your agent
⏯️Watch the demo: www.youtube.com/watch?v=z6l5nyE_6-I
Google for Developers
@googledevs
Build voice agents with visibility into every interaction ✨
Learn how @carodivittorio from @LangChain uses the Google ADK, the Gemini Live model, and LangSmith to trace voice agent behavior, inspect tool calls, review interruption events, analyze token costs, and more. -
LangChain reposted
Brace
@BraceSproul
OpenWiki supports `.openwikiignore` in 0.2.5!
Specify paths which should be ignored by OpenWiki runs and avoid wasting tokens on parts of your codebase which don't matter.
Try it out today:
npm install -g openwiki
Many thanks to Sohail Sarkar for contributing this! -
LangChain
@LangChain
Interrupt is the place to hear from industry leaders on whatʼs next for agents. interrupt.langchain.com
At Interrupt SF, we hosted speakers from @cisco, @linkedin, and more as well as @AndrewYNg + @levie
Coming to Interrupt NYC + London: @Modal, @AstraZeneca, @JPMorgan, @Linkedin, & more
Interested in attending? Get your tickets for Interrupt while spaces last. -
LangChain reposted
Harrison Chase
@hwchase17
sharing more about how we evaluate different agents we build internally
common themes (from this and future benchmarks):
- standardize on harbor
- skill for going from traces/raw data to harbor tasks -
LangChain
@LangChain
In 15 minutes, Robert Xu from the LangChain team breaks down:
✅ What Deep Agents actually are
✅ How the "harness" concept fits into the model + harness equation
✅ How Deep Agents relate to LangChain and LangGraph
✅ How to move Deep Agents into productionThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Interrupt returns this fall!
📍 NYC: September 24
📍 London: October 13
Spaces are limited. Get your tickets now! interrupt.langchain.com
LangChain
@LangChain
DB Engineering & Consulting is joining us at Interrupt London.
Gregor Beuster's team builds railway infrastructure plans, drawings too large for any AI to read at once.
Their coding agent builds and tests a vision agent, then reviews the traces in LangSmith to answer one question: what did the agent actually see?
Catch Gregor's talk + more @ Interrupt London, Oct 13. interrupt.langchain.com/london -
LangChain reposted
Nick Hollon
@nickhollon10
we built ReviewBench to measure how well our internal review agent was doing with respect to trusted human reviewers! @Vtrivedy10 eval engineering skill was tremendously helpful in speeding this process up!!
-
LangChain
@LangChain
We wanted a benchmark tied to the kinds of issues our reviewers catch in real PRs, so we built ReviewBench.
1️⃣Start from real reviews
2️⃣Curate them into concrete review issues
3️⃣Turn them into reproducible @harborframework tasks
An inside look from @nickhollon10. -
LangChain
@LangChain
DB Engineering & Consulting is joining us at Interrupt London.
Gregor Beuster's team builds railway infrastructure plans, drawings too large for any AI to read at once.
Their coding agent builds and tests a vision agent, then reviews the traces in LangSmith to answer one question: what did the agent actually see?
Catch Gregor's talk + more @ Interrupt London, Oct 13. interrupt.langchain.com/london -
LangChain reposted
Nebius
@nebiusai
We're building a compliance audit agent live on Tue 4th Aug: first on the Nebius Agents Blueprint, then with @LangChain Deep Agents on NVIDIA Nemotron 3 Ultra. Frontier-level quality at ~1/10 the cost, no fine-tuning.
Registration includes the recording. nebius.com/events/webinar-architecture-behind-production-… -
LangChain
@LangChain
✅ Avoid cost overruns
✅ Limit runaway traffic
✅ Improve agent reliability
✅ Reduce sensitive data exposure
One gateway across your models and providers.
LangChain
@LangChain
LangSmith LLM Gateway is now available in public beta.
Set spend and rate limits, determine fallback policies, and redact sensitive data before it reaches a model provider, all from one gateway that works across models and providers.
Learn more and try the public beta →
LangChain
@LangChain
More on LangSmith LLM Gateway: www.langchain.com/blog/langsmith-llm-gateway-runtime-cont…
-
LangChain
@LangChain
LangSmith LLM Gateway is now available in public beta.
Set spend and rate limits, determine fallback policies, and redact sensitive data before it reaches a model provider, all from one gateway that works across models and providers.
Learn more and try the public beta → -
LangChain reposted
Viv
@Vtrivedy10
New version of the eval-engineering skill released, thanks for everyone's feedback on the first version.
npx skills add github.com/langchain-ai/langchain-skills --skill eval-engineering
updates:
- now supports simulating users in multi-turn evals with instructions + examples
- more guidance on good Environment Design. examples on simulating backends/databases + how to use traces to make sure the env reflects prod
- more instructions on iteratively refining the harness, environment, and verifier with the user
Evals/Environments are hard, pls keep the feedback coming, more coming to reduce the friction of doing this while keeping domain experts in the loop! -
LangChain reposted
Oxylabs
@Oxylabs_io
Oxylabs x @LangChain
Here is how our official integration seamlessly feeds live web data into your LLM pipelines: -
LangChain reposted
Sydney Runkle
@sydneyrunkle
icymi -- we released deepagents v0.7 yesterday
it's more cost and token-efficient, more configurable, and has more performant filesystem tools
take it for a spin!
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
☑️ New from LangChain Academy: Become a LangChain Certified Agent Engineer
The first certification for the full Agent Development Lifecycle.
For the first 2 months, use code LCAE50 for 50% off. academy.langchain.com/pages/certifications-lcae -
LangChain
@LangChain
🎙️ Brand new Max Agency with @cognition President @russelljkaplan + @hwchase17.
⏯️ YouTube: youtu.be/bBUotstDLdk
🎧 Apple: podcasts.apple.com/us/podcast/the-misaligned-incentives-b…
🎧 Spotify: open.spotify.com/episode/7JNB5YXA6NVgUDAhy8eA3VThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Vinay Narayanamurthy (Principal Engineer, @HomeDepot) joins us at the LangSmith Roadshow in Atlanta on Aug 11.
He'll walk through how his team builds, tests, deploys, and monitors agents in production.
RSVP for keynotes, workshops, networking, and more:
events.langchain.com/LangSmithRoadshow/Atlanta/ -
LangChain
@LangChain
Don't miss our @ApacheDataFusio meetup with Hotdata on August 13th!
More info from @ankush_gola11 ⤵️
Ankush Gola
@ankush_gola11
We're really excited to be teaming up with our friends at Hotdata and hosting an @ApacheDataFusio meetup at our new SF office on Aug 13th!
The LangChain team will be giving a talk on how we built SmithDB.
We'll also have some great speakers: Divya Ranganathan, Alexander Bianchi, Shehab Amin, and Emil Sadek
If you're interested in modern data systems, you won't want to miss this. Register here: luma.com/gk3s2vib -
LangChain
@LangChain
How @Similarweb evaluates a Deep Research agent when there's no single right answer:
✅ Deterministic checks for tool calls
✅ Rubric-scored LLM judges for quality
✅ Faithfulness checks against retrieved data
✅ A/B comparisons against a saved baseline
All wired to traces in LangSmith.
www.langchain.com/blog/how-similarweb-evaluates-long-form… -
LangChain
@LangChain
OpenWiki now connects directly to LangSmith tracing projects to provide better context into how coding agents are actually interacting with your codebase.
To generate the most useful wikis you need to know where existing agents focus on the most, get stuck or fail. By giving OpenWiki direct access to LangSmith traces we can do exactly that.
Try OpenWiki locally today with just two commands:
npm install -g openwiki
openwiki --init
github.com/langchain-ai/openwiki
Brace
@BraceSproul
OpenWiki now connects to LangSmith traces to analyze how coding agents interact with your repo during wiki generations!
We added a LangSmith tracing connector so OpenWiki can retrieve more context into how coding agents (Claude Code, Codex, etc) interact with and write code in your repo.
You trace coding agent sessions to LangSmith (doc links in reply), then connect OpenWiki to those tracing projects so it can analyze where your agent is lacking context, where it gets stuck, and more to better design and write wikis.
Try the LangSmith tracing connector in OpenWiki today: github.com/langchain-ai/openwiki -
LangChain reposted
Brace
@BraceSproul
OpenWiki now connects to LangSmith traces to analyze how coding agents interact with your repo during wiki generations!
We added a LangSmith tracing connector so OpenWiki can retrieve more context into how coding agents (Claude Code, Codex, etc) interact with and write code in your repo.
You trace coding agent sessions to LangSmith (doc links in reply), then connect OpenWiki to those tracing projects so it can analyze where your agent is lacking context, where it gets stuck, and more to better design and write wikis.
Try the LangSmith tracing connector in OpenWiki today: github.com/langchain-ai/openwiki -
LangChain reposted
Sydney Runkle
@sydneyrunkle
We just released deepagents v0.7 with a leaner, more configurable harness! This release features:
- builtin middleware overrides (custom middleware prompts/settings)
- 65% reduction in the base prompt/tool descriptions
- optimized filesystem tools
Check it out!
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
Interrupt returns this fall!
📍 NYC: September 24
📍 London: October 13
Spaces are limited. Get your tickets now! interrupt.langchain.com
LangChain
@LangChain
.@UseApolloio will be at Interrupt NYC.
Apollo's AI Assistant is one of the earliest production multi-agent systems on LangGraph.
Their team will share how they migrated it from a hand-rolled supervisor to deepagents.create_deep_agent without breaking quality across production traffic.
Catch this talk + more @ Interrupt NYC, Sept 24. interrupt.langchain.com/nyc -
LangChain
@LangChain
"If I don't know what model I'm using, I say it's great. If I know it's the frontier model, my bias kicks in."
@FactoryAI CTO @enoreyes on why he thinks a lot of the model wars are just marketing.
LangChain
@LangChain
Check out the full conversation on the Max Agency podcast:
🎧 Apple: podcasts.apple.com/us/podcast/the-best-ai-agents-cost-les…
🎧 Spotify: open.spotify.com/episode/5CkX1dYgF89PGs4n3bJZFM
⏯️ YouTube: www.youtube.com/watch?v=HbUznYhKFOc -
LangChain
@LangChain
🎓 New LangChain Academy Course: Autonomous Agent Improvement with LangSmith Engine
In our latest course, we'll show you how to use Engine to:
✅ Identify and prioritize issues from traces
✅Draft fixes
✅Propose evals to prevent regressions
academy.langchain.com/courses/langsmith-engineThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
.@UseApolloio will be at Interrupt NYC.
Apollo's AI Assistant is one of the earliest production multi-agent systems on LangGraph.
Their team will share how they migrated it from a hand-rolled supervisor to deepagents.create_deep_agent without breaking quality across production traffic.
Catch this talk + more @ Interrupt NYC, Sept 24. interrupt.langchain.com/nyc -
LangChain
@LangChain
🚌 En route to improving your agents
-
LangChain
@LangChain
OpenWiki supports all the most common LLM providers.
Try it out today with one of the many we support:
github.com/langchain-ai/openwiki
Brace
@BraceSproul
OpenWiki supports 12 LLM providers:
- @OpenAI API & @ChatGPT sub login
- @AnthropicAI
- @GeminiApp AI Studio & Enterprise Vertex
- @awscloud Bedrock
- @OpenRouter
- @FireworksAI_HQ
- @baseten
- @nebiusai
- @nvidia NIM
So there's no excuse not to try it! (ChatGPT login makes this especially easy)
github.com/langchain-ai/openwiki -
LangChain reposted
Brace
@BraceSproul
OpenWiki supports 12 LLM providers:
- @OpenAI API & @ChatGPT sub login
- @AnthropicAI
- @GeminiApp AI Studio & Enterprise Vertex
- @awscloud Bedrock
- @OpenRouter
- @FireworksAI_HQ
- @baseten
- @nebiusai
- @nvidia NIM
So there's no excuse not to try it! (ChatGPT login makes this especially easy)
github.com/langchain-ai/openwiki -
LangChain
@LangChain
In our latest guide, we break down the four things every agent’s computer needs to do well: execute safely, stay in control, remain observable, and make it fast to build and iterate.
Read more: www.langchain.com/blog/agents-need-their-own-computer -
LangChain reposted
Brace
@BraceSproul
OpenWiki is still growing fast! Currently at 23.4k weekly downloads, 13.5k GitHub stars, almost 1k forks and over 50 unique contributors!
If you want easy and reliable docs for your codebase that automatically maintains itself, try it out today!
npm install -g openwiki
Repo: -
LangChain
@LangChain
Our data agent now handles roughly 40x the request volume our 3 person data team could manage directly.
Now, our data team can focus on the models, context, and guardrails that make the agent worth trusting in the first place.
How we built it:
www.langchain.com/blog/agent-data-stack -
LangChain
@LangChain
Coding is a general skill. Your company's domain intelligence is not.
Our partnership with @NVIDIA helps enterprises own and improve their domain intelligence instead of outsourcing it.
LangChain
@LangChain
Watch the full fireside chat between @hwchase17 + Jensen Huang: youtu.be/Yy3JH6dDugc
-
LangChain reposted
LangChain
@LangChain
.@JensenHuang + @hwchase17 fireside chat:
✅ Why companies need open agent systems
✅ NemoClaw for LangChain Deep Agents blueprintThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Headed to @Ai4Conferences in Vegas?
📍Visit us at Booth 1553
💻 8/3: Hands-on workshop on agent development
🏎️ 8/4: F1 Arcade event with @elastic + @googlecloud
(RSVP at events.elastic.co/f1arcadelasvegas)
🎤 8/6: Session on shipping agents in the enterprise -
LangChain
@LangChain
.@CreditGenie_US has debugged thousands of agent traces with LangSmith.
LangSmith traceability allows them to see exactly how their agent reasons, while LangSmith’s Insights feature helps them turn thousands of production traces into 100-200 targeted test questions per behavior.This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Happening tonight!
Sydney Runkle
@sydneyrunkle
boston friends -- @masondrxy and i are giving talks on loop eng and open models tonight in harvard square
come with good questions!
luma.com/iaie4h2x -
LangChain reposted
Sydney Runkle
@sydneyrunkle
use the latest and greatest open models w/ dcode!
Alex Olsen
@its_ao
GPT 5.6 -> Kimi K3 in 10 seconds with dcode
-
LangChain
@LangChain
Interrupt returns this fall!
📍 NYC: September 24
📍 London: October 13
Space are limited. Get your tickets now! interrupt.langchain.com
LangChain
@LangChain
At Interrupt, we’re running hands-on workshops hosted by our product experts.
1️⃣Deep Agents from Scratch
2️⃣Bringing Deep Agents into Production
3️⃣Autonomous Production Monitoring with LangSmith Engine
4️⃣Voice Agents (NYC Only)
Ready to level up? Don’t miss Interrupt NYC + London.
interrupt.langchain.com -
LangChain
@LangChain
👀
Sydney Runkle
@sydneyrunkle
new deepagents version is about to drop!
we just released a v0.7 beta:
1. much more minimal prompting -- we cut base input tokens by more than 1/2!
2. more configurability -- override any default middleware w/ your custom settings
test it out and let us know what you think! -
LangChain reposted
Viv
@Vtrivedy10
Open Questions to friends here:
1. Do you do evals?
2. What are friction point for building evals you actually trust?
3. What open tooling do you feel like you’re missing to make this easier?
im extending these skills + adding helper workflows so you can run everything locally/privately on your data
and would love to hear what people want
have heard:
- out of the box evaluators for tagging traces (that’s coming soon)
- support for simulations & multi-turn environments (coming)
- help with infra, harbor + sandboxes help here but happy to add more examples -
LangChain reposted
Alex Olsen
@its_ao
Kimi K3 x dcode!!
The open model moment just keeps happening
Mason Daugherty
@masondrxy
Available now in dcode!
docs.langchain.com/dcode
Fireworks
@FireworksAI_HQ
Kimi K3 is now live on Fireworks.
Get started → fireworks.ai/models/fireworks/kimi-k3 -
LangChain reposted
LangChain
@LangChain
Proud to be an inaugural partner alongside @NVIDIA and other industry leaders in the Open Secure AI Alliance.
We look forward to accelerating the adoption and trust of open models and harnesses through this initiative.
NVIDIA
@nvidia
AI security advances when the industry builds in the open, together.
We're introducing the Open Secure AI Alliance with industry leaders to develop new techniques and tools to safeguard software and agents.
By sharing models, tooling and research in the open, we can broaden the community of defenders.
Learn more about the founding members’ contributions: nvda.ws/4pD8Fc5 -
LangChain
@LangChain
Right before an AI cheats on a task, a "panic" can show up in its internals.
@FactoryAI CTO @enoreyes on how Factory's Droid is built to route around that failure mode.
LangChain
@LangChain
Check out the full conversation on the Max Agency podcast, hosted by @hwchase17.
🎧 Apple: podcasts.apple.com/us/podcast/the-best-ai-agents-cost-les…
🎧 Spotify: open.spotify.com/episode/5CkX1dYgF89PGs4n3bJZFM
⏯️ YouTube: www.youtube.com/watch?v=HbUznYhKFOc -
LangChain reposted
Mason Daugherty
@masondrxy
Available now in dcode!
docs.langchain.com/dcode
Fireworks
@FireworksAI_HQ
Kimi K3 is now live on Fireworks.
Get started → fireworks.ai/models/fireworks/kimi-k3 -
LangChain
@LangChain
At Interrupt, we’re running hands-on workshops hosted by our product experts.
1️⃣Deep Agents from Scratch
2️⃣Bringing Deep Agents into Production
3️⃣Autonomous Production Monitoring with LangSmith Engine
4️⃣Voice Agents (NYC Only)
Ready to level up? Don’t miss Interrupt NYC + London.
interrupt.langchain.com -
LangChain
@LangChain
ICYMI: Great read from @hwchase17
Harrison Chase
@hwchase17
-
LangChain
@LangChain
Proud to co-sign.
www.microsoft.com/en-us/corporate-responsibility/topics/o…
Satya Nadella
@satyanadella
Open-weight models are essential to a healthy AI ecosystem. Together with others across our industry, we are outlining a path for open-weight models to strengthen American competitiveness and expand economic opportunity, while protecting national security. www.microsoft.com/en-us/corporate-responsibility/topics/o…
-
LangChain
@LangChain
We just released Deep Agents v0.7.0b2 🚀
This release offers a more performant, more configurable base harness. Test it out and let us know what you think!
docs.langchain.com/oss/python/releases/changelog
Sydney Runkle
@sydneyrunkle
new deepagents version is about to drop!
we just released a v0.7 beta:
1. much more minimal prompting -- we cut base input tokens by more than 1/2!
2. more configurability -- override any default middleware w/ your custom settings
test it out and let us know what you think! -
LangChain
@LangChain
Proud to be an inaugural partner alongside @NVIDIA and other industry leaders in the Open Secure AI Alliance.
We look forward to accelerating the adoption and trust of open models and harnesses through this initiative.
NVIDIA
@nvidia
AI security advances when the industry builds in the open, together.
We're introducing the Open Secure AI Alliance with industry leaders to develop new techniques and tools to safeguard software and agents.
By sharing models, tooling and research in the open, we can broaden the community of defenders.
Learn more about the founding members’ contributions: nvda.ws/4pD8Fc5 -
LangChain reposted
Harrison Chase
@hwchase17
Every company that wants to build a business around AI will need to own their intelligence
As @JensenHuang @satyanadella and others wrote yesterday, a huge part of that is open models
But it’s more than that as well! This my take on what it means to own your intelligence
Harrison Chase
@hwchase17
-
LangChain
@LangChain
.@JensenHuang + @hwchase17 fireside chat:
✅ Why companies need open agent systems
✅ NemoClaw for LangChain Deep Agents blueprintThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
At LangChain, open source is in our DNA.
Open models bring more choice, more innovation, and more flexibility.
We're proud to support the leading open models through Deep Agents and the LangChain ecosystem.
Jensen Huang
@JensenHuang
For my first post, I’m sharing a letter @NVIDIA signed on why open models matter.
AI will transform every industry, power every company, and be built by every country.
Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.
The world needs both frontier closed models and frontier open models.
images.nvidia.com/pdf/Open-Weights-and-American-AI-Leader… -
LangChain
@LangChain
The @traversal_ai team sat down with @hwchase17 for a fireside chat on what it actually takes to build an AI SRE in production.
Watch the full session at youtu.be/U5PkKt_uJysThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
.@FactoryAI CTO @enoreyes ran the numbers.
Same code review task, wildly different price depending on the harness
Eno on why a a great model-agnostic harness can make any model better.
LangChain
@LangChain
Watch the full conversation on the Max Agency Podcast.
🎧 Apple: podcasts.apple.com/us/podcast/the-best-ai-agents-cost-les…
🎧 Spotify: open.spotify.com/episode/5CkX1dYgF89PGs4n3bJZFM
⏯️ YouTube: www.youtube.com/watch?v=HbUznYhKFOc -
LangChain
@LangChain
Agent governance extends beyond security, bringing together authentication, audit logs, user management, provider secrets, data separation, data residency, rate limits, fallbacks, and centralized spend controls.
On August 12th, we’ll share a practical framework for evaluating these requirements across cost, compliance, and operational scale, along with the infrastructure teams need as agents move from experimentation into production.
Save your spot: events.langchain.com/register/building-governed-agents/ -
LangChain reposted
Cognition
@cognition
DeepWiki has indexed over 500,000 repos. It's an easy way to get started when working with unfamiliar libraries, and free to use.
In this meetup with @LangChain and @jacobtpl, we'll share how it works under the hood.
LangChain
@LangChain
We’re hosting a meetup with @cognition on Tuesday.
Come learn about LLM Wikis and how to give your agents open memory with OpenWiki.
RSVP while spaces last: luma.com/mylwoabu -
LangChain reposted
Bridgewater Associates
@Bridgewater
Pocket Analyst Tool (PAT) is an example of how we're combining decades of codified investment knowledge with large language models, agentic workflows, and modern software architecture to build AI tools for investment research.
Watch how we built PAT.
LangChain
@LangChain
At Interrupt, the @Bridgewater team shared how they built the AIA Pocket Analyst Tool (PAT).
PAT is an internal AI analyst deployed to hundreds of investors that performs hours of deep exploratory research in minutes, leveraging Bridgewater’s proprietary data, methodologies, and expert investor feedback. PAT is one component of AIA, Bridgewater’s artificial investor.
Watch the keynote to see how they built it: youtu.be/lXZb21CfeIYThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
At Interrupt, the @Bridgewater team shared how they built the AIA Pocket Analyst Tool (PAT).
PAT is an internal AI analyst deployed to hundreds of investors that performs hours of deep exploratory research in minutes, leveraging Bridgewater’s proprietary data, methodologies, and expert investor feedback. PAT is one component of AIA, Bridgewater’s artificial investor.
Watch the keynote to see how they built it: youtu.be/lXZb21CfeIYThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Austin Vance
@austinbv
Hey I get interrupt on my birthday!
LangChain
@LangChain
Join us at Interrupt London on October 13th!
🎙️ @hwchase17, @astrazeneca, @roche, and more
💻 Technical sessions + hands-on workshops
🤝 Networking with other agent builders
🎟️ Tickets are limited. Save your spot at interrupt.langchain.com/london -
LangChain
@LangChain
Join us at Interrupt London on October 13th!
🎙️ @hwchase17, @astrazeneca, @roche, and more
💻 Technical sessions + hands-on workshops
🤝 Networking with other agent builders
🎟️ Tickets are limited. Save your spot at interrupt.langchain.com/london -
LangChain reposted
Sofia Montoya
@sofia_montoyac
Lucky to work with some of the best teams in AI - check out how @RilletHQ leverages LangSmith to monitor & improve their agents 💸
LangChain
@LangChain
LangSmith for Startups Spotlight: @RilletHQ
Rillet is the AI-native ERP built for mid-market and public companies.
✅ Powers 500+ customers like Mercor, Function Health, Kickstarter and more across AI, software, biotech, services, fintech, and beyond.
✅ Uses LangSmith as its observability platform to monitor and analyze its agents and surface inefficiencies.
Request a demo rillet.com/talk-to-sales -
LangChain
@LangChain
LangSmith for Startups Spotlight: @RilletHQ
Rillet is the AI-native ERP built for mid-market and public companies.
✅ Powers 500+ customers like Mercor, Function Health, Kickstarter and more across AI, software, biotech, services, fintech, and beyond.
✅ Uses LangSmith as its observability platform to monitor and analyze its agents and surface inefficiencies.
Request a demo rillet.com/talk-to-sales -
LangChain
@LangChain
.@Salesforce's Agentforce Vibes team just crossed 100M lines of accepted code.
LangSmith is how they know it's good enough to ship.This media couldn't be displayed. See it on the original post. -
LangChain reposted
Sydney Runkle
@sydneyrunkle
going live in 20 mins to talk about all things loop engineering!
come join!
events.langchain.com/webinar/the-art-of-loop-engineering/ -
LangChain reposted
Philipp Schmid
@_philschmid
If you are building real-time voice agents with @GoogleDeepMind Gemini Live, you can now trace your speech-to-speech agent loops directly in @LangChain!
- Speaker callback hooks capture only the exact audio your user heard before an interrupted event.
- Live timeline highlighting user speech in blue, Gemini responses in orange, and custom tool executions (like a 1.29s city weather lookup).
- Usage metadata separates regular text tokens from native Gemini audio tokens so you always know your exact usage footprint.
To try it out. In your Google ADK agent, register LangSmithGoogleADKLivePlugin and done.
[Video Demo by Caroline di Vittorio]This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
An inside look from @nickhollon10 ⤵️
LangChain OSS
@LangChain_OSS
-
LangChain
@LangChain
.@sydneyrunkle on the common applications for Deep Agents.
LangChain
@LangChain
See more in our Deep Agents Fireside Chat.
youtu.be/GbzEDgcuGJU -
LangChain reposted
Brace
@BraceSproul
OpenWiki has Mermaid diagram support now!
When generating wikis it's automatically prompted to also design diagrams to include in the docs. This is one small step towards building a wiki experience that's not only great for agents, but also for us humans too!
Checkout OpenWiki today (fully OSS wiki generator CLI) and let us know what you think 👇 -
LangChain
@LangChain
Agent governance extends beyond security, bringing together authentication, audit logs, user management, provider secrets, data separation, data residency, rate limits, fallbacks, and centralized spend controls.
In Building Governed Agents, we’ll share a practical framework for evaluating these requirements across cost, compliance, and operational scale, along with the infrastructure teams need as agents move from experimentation into production.
Save your spot: events.langchain.com/register/building-governed-agents/ -
LangChain
@LangChain
Lower inference costs makes running and evaluating more specialized agents in production more practical.
Teams can create agents for specific domains, use evals and traces to measure performance, and adapt the harness as their workflows change.
LangChain
@LangChain
Watch the full fireside chat between @hwchase17 + Jensen Huang.
www.youtube.com/watch?v=Yy3JH6dDugc -
LangChain reposted
LangChain
@LangChain
.@its_ao added native browser control to dcode with one /goal.
Here’s how you can use /goal to do the same: docs.langchain.com/oss/python/deepagents/code/overviewThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
The difference between an agent that works once and one that keeps improving often comes down to the loops around it, including how it checks its work, responds to new inputs, and turns production traces into better prompts, tools, graders, and configurations.
Join us for The Art Of Loop Engineering to learn how these loops fit together, where human oversight matters most, and how LangChain and LangSmith support the full improvement cycle.
Register before the session: events.langchain.com/webinar/the-art-of-loop-engineering/ -
LangChain reposted
Sydney Runkle
@sydneyrunkle
open evals! @Vtrivedy10 and our labs team are making it easy for you to automate quality eval gen
this is an essential part of continual learning for agents
eval loops, people!!! -
LangChain
@LangChain
We’re hosting a meetup with @cognition on Tuesday.
Come learn about LLM Wikis and how to give your agents open memory with OpenWiki.
RSVP while spaces last: luma.com/mylwoabu -
LangChain
@LangChain
Join us at Interrupt London on October 13th!
🎙️ @hwchase17, @jpmorgan, @useapolloio, @modal, and more
💻 Technical sessions + hands-on workshops
🤝 Networking with other agent builders
🎟️ Tickets are limited. Save your spot at interrupt.langchain.com/nyc -
LangChain
@LangChain
Available today in the langchain-ai/langchain-skills repo: github.com/langchain-ai/langchain-skills
💻 Install in Codex or Claude Code, open the repo w/ the agent you want to evaluate, start with a prompt
✨ Get a @Harborframework task under evals/, an actual target run, and a review of whether the verifier measured the intended behavior correctly. -
LangChain reposted
Viv
@Vtrivedy10
democratizing high quality evals & environments for your specific use-cases is a huge win for Open Intelligence.
so we’re sharing skills & workflows that plug right into your coding agent to help everyone own this process using their Trace data using the open @harborframework
A design principle we adopt is: The best way to make good evals is to interview the user about what the agent should be good at!
Evals are training data for agents. The goal of evals is to transfer knowledge and specialization encoded in Tasks into Agent behavior. A lot of that knowledge lives in human minds, and interviewing them and iteratively testing evals is an effective way to drive that transfer.
Creating evals is hard. People are unfamiliar with the tooling and goals are often underspecified in a first pass. Agent optimization is iterative because you don’t know what agents will do until you run them. The same holds for evals which is why we find that iteratively refining them is usually much better than one-shot generation.
A large part of our research mission towards Continual Learning revolves around large-scale data mining and good eval creation. Every team should be able to use their data to continuously improve their agents
looking forward to getting community feedback on this flow, refining it, releasing more data, and helping more users own their intelligence :) -
LangChain reposted
-
LangChain
@LangChain
We’re launching the Eval Engineering Skill, a skill that helps coding agents build evals using context from a repository + agent traces.
Everything you need to know from @vtrivedy10 ⤵️ -
LangChain reposted
Alex Olsen
@its_ao
Loops are everywhere!
Maybe the bigger the loop, the more meaningful work you can do
watch dcode nail a huge feature in one shot with /goal
LangChain
@LangChain
.@its_ao added native browser control to dcode with one /goal.
Here’s how you can use /goal to do the same: docs.langchain.com/oss/python/deepagents/code/overviewThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Every turn of a conversation has two pieces.
👤 What the user did: observable signals that tell you where the conversation is
🤖 What the agent did: controllable inputs that condition where it goes next
@GetCandidly's IO-HMM keeps these separate by design.
👤 User behavior reads the state
🤖 Agent behavior shifts it
LangChain
@LangChain
Full story: www.langchain.com/blog/how-candidly-built-state-aware-age…
-
LangChain
@LangChain
.@its_ao added native browser control to dcode with one /goal.
Here’s how you can use /goal to do the same: docs.langchain.com/oss/python/deepagents/code/overviewThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Interrupt, The Agent Conference by LangChain returns this fall:
📍NYC: September 24
📍London: October 13
Get more info + save your seat while spaces last at interrupt.langchain.com
LangChain
@LangChain
At Interrupt, we learned how LinkedIn cut time-to-interview by 60% By building a hiring agent with LangGraph.
Watch the full session: youtu.be/LUemJGG2k4cThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Benjamin Tannyhill
@bentannyhill
Great walkthrough of how we built/run evals for Engine!
-
LangChain
@LangChain
At Interrupt, we learned how LinkedIn cut time-to-interview by 60% By building a hiring agent with LangGraph.
Watch the full session: youtu.be/LUemJGG2k4cThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
"Graph engineering" is having a moment, but it's not a new thing!
We've been building agents powered by graphs for three years, here's what we've learned!
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
.@useapolloio cut dev-to-launch work for new AI Assistant skills by 80-85%, by tearing out a hand-wired supervisor for Deep Agents.
Full story: www.langchain.com/blog/how-apollo-rebuilt-its-ai-assistan… -
LangChain reposted
NVIDIA AI
@NVIDIAAI
Tune the Harness, Before Tuning the Model with LangChain | Nemotron Labs x.com/i/broadcasts/1MJgNNeWbklGL
-
LangChain reposted
Sofia Montoya
@sofia_montoyac
Huge launch from our team!! If you're building voice agents, reach out to learn more about how you can trace them in LangSmith
LangChain
@LangChain
Voice agents are exploding. Don’t let them be a black box in production.
Today, we’re launching LangSmith tracing for 4 voice frameworks:
🎙️ @pipecat_ai
🎙️ @livekit
🎙️ @OpenAI Realtime
🎙️ @GeminiApp Live (Google ADK)
Learn more: www.langchain.com/blog/trace-voice-agents-in-langsmith -
LangChain
@LangChain
🎧 Full audio on your traces
⚡ STT/TTS latency
🗣️ interruptions + VAD
Only a few lines of code to set up👇
LangChain
@LangChain
Voice agents are exploding. Don’t let them be a black box in production.
Today, we’re launching LangSmith tracing for 4 voice frameworks:
🎙️ @pipecat_ai
🎙️ @livekit
🎙️ @OpenAI Realtime
🎙️ @GeminiApp Live (Google ADK)
Learn more: www.langchain.com/blog/trace-voice-agents-in-langsmith -
LangChain
@LangChain
Google Gemini AI Studio, Enterprise Vertex and the new Gemini 3.6 Flash, and 3.5 Flash Lite now available in OpenWiki
Brace
@BraceSproul
OpenWiki now supports Gemini AI Studio & Enterprise Vertex AI!
This change also adds the new Gemini 3.6 Flash and 3.5 Flash Lite (released today)
Big thank you to @bradhuffman and @sadrig91 for contributing these two PRs!
Try it out today, it's fully OSS and the easiest way to generate and maintain codebase documentation 👇 -
LangChain reposted
Brace
@BraceSproul
OpenWiki now supports Gemini AI Studio & Enterprise Vertex AI!
This change also adds the new Gemini 3.6 Flash and 3.5 Flash Lite (released today)
Big thank you to @bradhuffman and @sadrig91 for contributing these two PRs!
Try it out today, it's fully OSS and the easiest way to generate and maintain codebase documentation 👇 -
LangChain
@LangChain
Voice agents are exploding. Don’t let them be a black box in production.
Today, we’re launching LangSmith tracing for 4 voice frameworks:
🎙️ @pipecat_ai
🎙️ @livekit
🎙️ @OpenAI Realtime
🎙️ @GeminiApp Live (Google ADK)
Learn more: www.langchain.com/blog/trace-voice-agents-in-langsmith -
LangChain
@LangChain
We built a tracing plugin that turns every @cursor_ai agent session into a structured trace in LangSmith: the model run, every tool call underneath it, and any nested subagent work.
Attachments get recovered and rendered inline too.
This is the third video in our coding-agent tracing series (Claude Code, Codex, Cursor). All three share the same trace schema, so you can compare them in one place.This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
"Once you've found the shape of intelligence, you know what it looks like."
@enoreyes from @factoryAI on why he thinks distillation is basically unstoppable and why model labs might already know it.
LangChain
@LangChain
Watch the full conversation on the Max Agency podcast:
⏯️ YouTube: www.youtube.com/watch?v=HbUznYhKFOc
🎧 Apple: podcasts.apple.com/us/podcast/the-best-ai-agents-cost-les…
🎧 Spotify: open.spotify.com/episode/5CkX1dYgF89PGs4n3bJZFM -
LangChain
@LangChain
Interrupt, The Agent Conference by LangChain returns this fall with two new stops:
📍NYC: September 24
📍London: October 13
Get more information about Interrupt + save your seat while spaces last at interrupt.langchain.comThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
The Agent Development Lifecycle is continuous building, deploying, monitoring, improving, and governing agents as they move into production.
Join the next sessions in Build More with LangSmith: Summer AMA Series to learn how LangSmith supports more of that lifecycle:
✅ July 23 — Build More with LangSmith Fleet
✅ July 30 — Build More with LangSmith Insights
Explore the series and save your spot: events.langchain.com/ama-series/build-more-with-langsmith/ -
LangChain reposted
LangChain
@LangChain
We’re making LangSmith Sandboxes available to try for free.
Start today: smith.langchain.comThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Viv
@Vtrivedy10
Agents are reflections of the data they’re fit on! And to make Engine good at mining massive scale, diverse Trace data, we had to synthetically generate data that reflected those domains
this includes settings like varying Trace lengths, allowing user customization for what they care about, and experimenting with different agent architectures for parsing data
this gives a signal for us to hill climb against for model choice and discovering better harness design
lots of iterative engineering goes into building a bench! “who evaluates the evaluator?” -> us!
give it a read! more coming here -
LangChain reposted
Harrison Chase
@hwchase17
LangSmith Engine is our in product agent that runs over traces, clusters them into issues, and proposes fixes
It's a long running, complex, ambiguous process
We wrote about how we evaluate it! -
LangChain
@LangChain
IssueBench is a detailed evaluation suite we built for Engine (continual learning agent in LangSmith, runs on your traces, automatically improves your agent)
Evaluating Engine was a tricky problem due to the type of data it needs to run on, so we wrote this blog outlining what the benchmark is, and how we made it: -
LangChain reposted
-
LangChain
@LangChain
Governing agents requires teams to think across the full system, from authentication, audit logs, rate limits, fallbacks, and centralized spend controls.
Our latest conceptual guide breaks down the foundations teams need.
www.langchain.com/blog/building-governed-agents-a-framewo… -
LangChain
@LangChain
We’re making LangSmith Sandboxes available to try for free.
Start today: smith.langchain.comThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Deep Agents explained in one sentence by @sydneyrunkle
LangChain
@LangChain
See more in our Deep Agents Fireside Chat.
youtu.be/GbzEDgcuGJU -
LangChain
@LangChain
Interrupt returns this fall!
📍 NYC: September 24
📍 London: October 13
Info + RSVP: interrupt.langchain.com
LangChain
@LangChain
At Interrupt, @Coinbase Engineering Manager Evan Kormos took the stage to share how their developer support team built a self-improving agentic system on top of self-hosted LangSmith.
Full session: youtu.be/py9d6zTl4Dc -
LangChain
@LangChain
At Interrupt, @Coinbase Engineering Manager Evan Kormos took the stage to share how their developer support team built a self-improving agentic system on top of self-hosted LangSmith.
Full session: youtu.be/py9d6zTl4Dc -
LangChain
@LangChain
We've built a full software engineering agent factory in house, and open-sourced every piece of it. @BraceSproul's blog dives into each of those components, what they're used for, and why them being OSS matters
-
LangChain reposted
-
LangChain
@LangChain
GPT Researcher, built on deepagents, is the easiest way to get started building a deep research agent!
Sydney Runkle
@sydneyrunkle
GPT Researcher from @assaf_elovic and @tavilyai is a high-performance, open-source deep research implementation (almost at 30k ⭐), now built on @LangChain's Deep Agents.
It uses the same plan → parallel research → review → publish pipeline as the original LangGraph implementation, but with far less code: planning, delegation & context management are emergent in the harness rather than hardcoded in the graph.
Swapping GPT researcher in for tavily web search gives ~2x the verified citations at equal-or-better quality, with GPT-researcher winning in 7/10 Deep Research Bench tasks!
github.com/assafelovic/gpt-researcher/blob/main/deep_agen… -
LangChain
@LangChain
👀
Brace
@BraceSproul
i need to hire someone who's really really good at evals and benchmarks in SF (or maybe nyc if you're super special)
if this is you, or you know someone who fits this, please dm me! i'd love to chat
LangChain
@LangChain
LangChain is hiring.
Check out our open roles and apply
www.langchain.com/careers -
LangChain
@LangChain
This agent turns an RFP into a fully drafted response, citations included, letting the SME review start from a draft instead of a blank page.
1️⃣ Requirements extraction agent: Pulls scope, SLAs, pricing format, + regulatory provisions from RFP package
2️⃣ Content mapping agent: Matches requirements to approved internal content
3️⃣ Draft generation agent: Writes responses with citations
4️⃣ Gap detection agent: Flags exactly what needs a human
Built with LangGraph + LangSmith, measured with @payicorp. www.langchain.com/blog/proving-the-roi-of-agentic-ai-in-f… -
LangChain
@LangChain
In 40 minutes, @sydneyrunkle and @jakebroekhuizen run through everything you need to know about Deep Agents.
What are Deep Agents?
How do they work?
What are some of the most common use cases?
What are the core components?
How do file systems work?
Watch the full session ⤵️This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Agents are a new kind of software: endless inputs, non-deterministic outputs
Build them like normal software and they break
We built LangSmith around the Agent Development Lifecycle:
Build → Test → Deploy → Monitor
The lifecycle explained ⤵️This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
On July 23rd, @sydneyrunkle will break down how teams can use agent, verification, event-driven, and hill-climbing loops to build systems that are more observable, testable, and dependable over time.
Save your spot: events.langchain.com/webinar/the-art-of-loop-engineering/ -
LangChain
@LangChain
"Trust is earned in drops, but lost in buckets.”
@AbridgeHQ's VP of Product Janie Lee on what it actually takes to ship AI in clinical settings.
Watch the full Interrupt keynote: youtu.be/mxweSHetuN8
LangChain
@LangChain
Interrupt returns this fall
📍 NYC: September 24
📍 London: October 13
Save your spot ⤵️
interrupt.langchain.com -
LangChain
@LangChain
We’re hosting a meetup with @cognition on July 28th.
Come learn about LLM Wikis and how to give your agents open memory with OpenWiki.
RSVP while spaces last: luma.com/mylwoabu -
LangChain reposted
Factory
@factory
Our CTO @EnoReyes sat down with Harrison Chase to talk models, harnesses, and what it actually takes to build a software factory.
Harrison Chase
@hwchase17
New Max Agency with @FactoryAI CTO & Co-Founder @EnoReyes.
This was a particularly fun one. Since we're both building harnesses, we nerded out pretty deep.
Awesome conversation on why the harness matters more than the model running underneath it, how Factory built Missions, + more.
⏯️ YouTube: www.youtube.com/watch?v=HbUznYhKFOc
🎧 Apple: podcasts.apple.com/us/podcast/the-best-ai-agents-cost-les…
🎧 Spotify: open.spotify.com/episode/5CkX1dYgF89PGs4n3bJZFMThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Colin Francis
@colifran_
OpenWiki 0.2 adopts OKF!
this is a big deal for two reasons:
1. OKF adds structured metadata to every doc and builds connections through cross links which means faster results, fewer tokens
2. the growing OKF tools ecosystem lets you visualize your wiki and the connections between docs -
LangChain
@LangChain
As agents take on larger workloads across the organization, governance becomes part of the system design, shaping how teams control spend, protect sensitive data, enforce internal policies, and manage access at scale.
On August 12th, we’ll break down the foundations teams need to govern production agents and where gateway infrastructure fits into the broader Agent Development Lifecycle.
events.langchain.com/register/building-governed-agents/ -
LangChain reposted
Sydney Runkle
@sydneyrunkle
owning intelligence doesn't mean being able to afford the best model
it means owning the agent development lifecycle from model agnostic harness -> observability -> learning from each run
Harrison Chase
@hwchase17
what does it really mean to "own your intelligence"?
1. can the system compound *on your terms*? the durable advantage is not the model/system at this point in time - it's the flywheel that compounds over time
2. intelligence is no longer just the model - its the system around the model. if the system is opaque or locked away, you are renting intelligence (not owning it)
3. own the harness. integrate it deeply into your systems. make it specific to the task at hand
4. own the context and memory layer. includes user preferences, org knowledge, workflow patterns. if memory is trapped in vendor product, your learning is trapped to
5. own model optionality. you need freedom to route across models, providers, deployment modes. different tasks require different tradeoffs
6. own the economics. intelligence only matters if you can afford to deploy it broadly. sustainable intelligence at scale
7. own the feedback loop. observability tells you what happened, evals tell you if it worked. without this loop you cannot measure the intelligence, you are guessing
s/o @nlarusstone @nfcampos for their thoughts on feedback -
LangChain reposted
Sydney Runkle
@sydneyrunkle
come join us in boston to chat about loop engineering and open models!
luma.com/iaie4h2x -
LangChain reposted
Jacob Lee
@Hacubu
This year @LangChain 🦜 🔗 has almost tripled in size, from ~100 people to over 300! That's a lot of new faces, and I'm often asked: "What made you join early? What was the company like at seven people?"
I'm always honest: I wasn't convinced right away. I mean, three years ago, the website looked like this: web.archive.org/web/20230303225335/https://langchain.com
Raw HTML and two emojis for a logo. When I told my parents I was joining, they asked, pretty directly, if I'd lost my mind.
I met @ankush_gola11 about a month after ChatGPT launched and the open-source started getting real traction. I was completely fascinated by the technology, but my reaction to his pitch was "cool - good luck!" In addition to said website, I had just come off six years founding a company, and didn't want to jump back into an early startup.
A few weeks later, a client wanted to build an LLM-powered extraction pipeline - nowhere near as trivial then as it is now. I tried an early version of LangChain.js, and it worked great, so I kept exploring and found the section on agents.
The magic there completely hooked me. It truly seemed like something out of science fiction.
Thankfully, @hwchase17 and Ankush were still looking to round out their founding team, and I got to see how they operated up close. Unlike most startups, LangChain has never had a name recognition problem, and I knew firsthand how easy it was to start chasing GitHub stars and other vanity metrics.
But from the start, the team was serious about seeing through the fluff, listening to the community even when tempers flared, and building what actually solved user problems. They weren't always right - I joke that back then we were a six-month-old startup with a year of tech debt - but I admired how quickly they'd throw out popular beliefs and abstractions the moment evidence showed they were wrong or wouldn't scale. Anyone remember ConversationalRetrievalChain?
The website itself was emblematic of this "zero tolerance for bullshit" approach (as Harrison puts it). The small early team could have spent time creating a beautiful marketing site, but marketing wasn't the problem back then (though we're hiring now!). A parrot and a link emoji was enough. Time was better spent shipping.
So, what was it actually like at seven people? Different principles, but ones that rhyme with the old ones. "Zero tolerance for bullshit" has become "act with maximum agency". Our most interesting Slack discussions start with "Hot take", and it's a mark of dishonor if people react with ice cubes, meaning it's not extreme enough to encourage a rethink of a longstanding belief.
But the north star hasn't moved: build something great, and cut everything that isn't that.
There's a lot of story left to write. We're hiring across product, engineering, office management, people ops, and more! If that sounds like a place you'd thrive, come find us.
P.S. the OG site is a lot more clever than it looks - see www.berkshirehathaway.com -
LangChain reposted
Brace
@BraceSproul
I just published a short YouTube video on OKF in OpenWiki. It's only 3:30 long, so if you're curious what OKF is, what it'll mean for OpenWiki, and why we decided to adopt, you should give it a watch!
youtu.be/NxJjMvDN6aE -
LangChain
@LangChain
OpenWiki is adopting OKF in 0.2!
OKF will lay the ground work for a lot of new exciting features and capabilities in OpenWiki, and it'll allow us to take advantage of the growing OSS ecosystem around OKF!
Checkout @BraceSproul's short blog on what this means for OpenWiki, and its users: -
LangChain reposted
-
LangChain reposted
Brace
@BraceSproul
Super excited to officially adopt OKF in OpenWiki! There's a ton of exciting things this will allow for & make easier. I wrote a super short (~5min read) blog on what this will means for codebase memory in OpenWiki
Tldr;
- better structure of wikis
- foundation for search & retrieval tools (not just agentic search)
- a strong oss ecosystem already -
LangChain
@LangChain
New video from @its_ao.
How to run @NVIDIA Nemotron 3 Ultra with @baseten + Deep Agents Code.
✅ 550B parameters, up to 300 tokens per second
✅ Terminal agent w/ skills, sub-agents, MCP support
✅ First-class tracing through LangSmithThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Brand new Max Agency with @FactoryAI CTO & Co-Founder @EnoReyes!
Harrison Chase
@hwchase17
New Max Agency with @FactoryAI CTO & Co-Founder @EnoReyes.
This was a particularly fun one. Since we're both building harnesses, we nerded out pretty deep.
Awesome conversation on why the harness matters more than the model running underneath it, how Factory built Missions, + more.
⏯️ YouTube: www.youtube.com/watch?v=HbUznYhKFOc
🎧 Apple: podcasts.apple.com/us/podcast/the-best-ai-agents-cost-les…
🎧 Spotify: open.spotify.com/episode/5CkX1dYgF89PGs4n3bJZFMThis media couldn't be displayed. See it on the original post.
LangChain
@LangChain
🎧 Apple: podcasts.apple.com/us/podcast/the-best-ai-agents-cost-les…
🎧 Spotify: open.spotify.com/episode/5CkX1dYgF89PGs4n3bJZFM
⏯️ YouTube: www.youtube.com/watch?v=HbUznYhKFOc -
LangChain
@LangChain
How @Box uses Deep Agents middleware for Box Agent:
1️⃣Citation generation: Streaming of the answer + citation generation happens in parallel to avoid user interruption
2️⃣Prompt caching: Injects caching on multi-turn conversations, reducing cost + latency
3️⃣Context management: Summarizes conversation history after exceeding 170k tokens, preventing context overflow
www.langchain.com/blog/building-box-ai-how-an-enterprise-… -
LangChain
@LangChain
Governing agents requires teams to think across the full system, from authentication, audit logs, rate limits, fallbacks, and centralized spend controls.
Our latest conceptual guide, breaks down the foundations teams need to manage access, protect sensitive data, enforce policies, and control costs as agents move into production.
www.langchain.com/blog/agents-need-their-own-computer -
LangChain reposted
Brace
@BraceSproul
Most agents fail because they don't have the right context, but because the model is weak.
That's the problem LLM wikis are trying to solve: a structured knowledge base an agent can query to understand a codebase, product, or domain. We recently open sourced OpenWiki, our take on building and maintaining one of these.
We're hosting a meetup in our new office to dig into this pattern with @jacobtpl from @cognition's research and engineering team.
I'll walk through how we actually build these systems, and push on whether "wiki" is even the right frame for what they're doing. Then Jacob will share how Cognition built DeepWiki.
If you're building agents and thinking about context, come argue with us: luma.com/mylwoabu -
LangChain
@LangChain
ICYMI
LangChain
@LangChain
Fleet in Slack just got a big upgrade.
Now you can bring any Fleet agent into Slack in one click. Give it a custom identity. Use it in channels and threads. Hand off files, approve its actions, and keep all the context of your work in one place.
Read more in our blog: -
LangChain
@LangChain
LangSmith Roadshow:
✅ A half-day, hands-on look at LangSmith Engine + Managed Deep Agents
✅ Live demos, customer stories, lab sessions, and more
LangChain
@LangChain
Upcoming stops:
📍8/11: Atlanta
📍8/27: Silicon Valley
📍9/29: Boston
Get more info + RSVP for a LangSmith Roadshow event near you:
events.langchain.com/LangSmithRoadshow/ -
LangChain reposted
Sydney Runkle
@sydneyrunkle
i'm running a webinar next week on the art of loop engineering! come join to learn about agents that automatically improve over time
events.langchain.com/webinar/the-art-of-loop-engineering/ -
LangChain
@LangChain
.@11x_official started with one Fleet agent for bug triage.
Soon, they had agents for on-call alerts and product Q&A for GTM.
The reason adoption spread: every agent had a clear identity in Slack.
“Everyone at the company knows the identity of the agent, and it’s in all the main channels, so they just tag it. We even have salespeople who pull it up on sales calls when they have a question about the product.”
Fleet agents can now be deployed to Slack in one click.This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
.@caspar_br on how we're making it simple to equip your team with custom agents in Slack, without writing code.
-
LangChain reposted
Brace
@BraceSproul
The team has been pushing a ton to improve Fleet in every area! We're doubling down on Slack (based on how everyone internally primarily uses Fleet agents)
It's now easier than ever to create agents, give them their own Slack identities, and use in the place you already work
LangChain
@LangChain
Fleet in Slack just got a big upgrade.
Now you can bring any Fleet agent into Slack in one click. Give it a custom identity. Use it in channels and threads. Hand off files, approve its actions, and keep all the context of your work in one place.
Read more in our blog: -
LangChain reposted
Alex Olsen
@its_ao
I can't understate how big of an unlock this is
Early single-agent/multi-tenant startups were really cool, but this is so much bigger
LangChain
@LangChain
Fleet in Slack just got a big upgrade.
Now you can bring any Fleet agent into Slack in one click. Give it a custom identity. Use it in channels and threads. Hand off files, approve its actions, and keep all the context of your work in one place.
Read more in our blog: -
LangChain reposted
Harrison Chase
@hwchase17
It’s super easy to create custom slack agents now
Makes it easy to bring intelligence to where the team already lives
LangChain
@LangChain
Fleet in Slack just got a big upgrade.
Now you can bring any Fleet agent into Slack in one click. Give it a custom identity. Use it in channels and threads. Hand off files, approve its actions, and keep all the context of your work in one place.
Read more in our blog: -
LangChain
@LangChain
Fleet in Slack just got a big upgrade.
Now you can bring any Fleet agent into Slack in one click. Give it a custom identity. Use it in channels and threads. Hand off files, approve its actions, and keep all the context of your work in one place.
Read more in our blog: -
LangChain reposted
Hari
@harisaiharish
my first @langchain article! more to come soon (hopefully)👀
LangChain
@LangChain
Cursor, Copilot, Pi, and OpenCode tracing: Now in LangSmith.
Full session observability, no extra instrumentation.
✅Identify, group, and query any coding-agent trace with the same stable keys, regardless of which agent produced it
✅See the full run tree: turns, model calls, tools, and subagents
✅Token usage and cost per session, out of the box
Blog by @harisaiharish
www.langchain.com/blog/your-coding-agents-are-a-black-box… -
LangChain
@LangChain
LangSmith Engine originally auto-created PRs for every issue found,. This was too noisy for users.
@BenTannyhill on why switching to an inbox model that surfaces clustered problems with history became the better design.
LangChain
@LangChain
Listen to the full conversation on your favorite podcasting platform.
⏯️ YouTube: youtu.be/YqjR4vQwbTc
🎧 Apple: podcasts.apple.com/nz/podcast/the-best-ai-agents-are-secr…
🟢 Spotify: open.spotify.com/episode/5lT8hva20sVmJ0WiVCLAaO -
LangChain
@LangChain
Happening today @ 11am PT / 2pm ET!
Learn how to safely give agents their own computer.
events.langchain.com/register/build-a-secure-computer-for… -
LangChain
@LangChain
Interrupt is hitting the road this fall:
📍 NYC- September 24
📍 London- October 13
Info + RSVP: interrupt.langchain.com
LangChain
@LangChain
At Interrupt, @RCUmmadisetti and @kordelfrance from @Toyota's enterprise AI team took us behind the scenes of ToyotaGPT, a LangGraph-powered platform that reduced agent delivery from 6 months to 4 days, with over 50 agents now in production. youtu.be/nUNuNxMhwug
-
LangChain
@LangChain
At Interrupt, @RCUmmadisetti and @kordelfrance from @Toyota's enterprise AI team took us behind the scenes of ToyotaGPT, a LangGraph-powered platform that reduced agent delivery from 6 months to 4 days, with over 50 agents now in production. youtu.be/nUNuNxMhwug
-
LangChain
@LangChain
The most searched questions about agents, answered by the LangChain team.
This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Build more with LangSmith this summer.
Join our Summer AMA Series for practical walkthroughs + live Q&A with the LangChain team.
events.langchain.com/ama-series/build-more-with-langsmith/ -
LangChain
@LangChain
Cursor, Copilot, Pi, and OpenCode tracing: Now in LangSmith.
Full session observability, no extra instrumentation.
✅Identify, group, and query any coding-agent trace with the same stable keys, regardless of which agent produced it
✅See the full run tree: turns, model calls, tools, and subagents
✅Token usage and cost per session, out of the box
Blog by @harisaiharish
www.langchain.com/blog/your-coding-agents-are-a-black-box…
LangChain
@LangChain
Get started today.
Docs: docs.langchain.com/langsmith/integrations -
LangChain
@LangChain
We built a tracing plugin for every Codex session into LangSmith. Now every turn (tool calls, token usage, subagent threads) lands in LangSmith as a real trace you can dig into.
Two config blocks and one flag, and it's live.This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Building an agent that can take action is only the first step, because reliable systems also need loops that help them complete tasks, verify their work, respond to real-world events, and improve from what happens in production.
On July 23rd, join @sydneyrunkle for a practical session on the four loops behind production-ready agents and how they work together.
events.langchain.com/webinar/the-art-of-loop-engineering/ -
LangChain reposted
Viv
@Vtrivedy10
what a cool message to see :)
3 years ago when LLMs blew up, @LangChain was pretty much the only thing out there to build something cool and get a “wow” moment ✨
a chat app, a RAG app, the first agents, it was all there moving 1 million miles an hour
the dev community around ai has grown a ton since those days with new models, tooling, primitives, etc
but it’s still the same awesome feeling seeing builders like Eshwar & others in the community diving into the tooling and making cool stuff
that’s the magic of getting to work with a great open source team & community
if we can help, hit us up, the OSS team here is amazing and wants to help ❤️
Eshwar Prasad Yaddanapudi
@yesprasad26
After office hours - building Agents with @LangChain #langgraph . There’s still a couple weeks before I finish this and go into deep agents. Really appreciate the efforts from @LangChain_OSS team @Vtrivedy10 and team for everyday inspiration and insights
-
LangChain
@LangChain
Jensen Huang on what changed in the last six months.
Agentic systems grounded in knowledge, with tools, memory, and the ability to iterate until the job gets done.
Now, the models have finally caught up to make it work.
LangChain
@LangChain
ICYMI: www.langchain.com/blog/langchain-and-nvidia-launch-the-ne…
-
LangChain
@LangChain
Product Manager @BenTannyhill on how our product engineering team ships fast.
LangChain
@LangChain
Listen to the full conversation on your favorite podcasting platform.
⏯️ YouTube: youtu.be/YqjR4vQwbTc
🎧 Apple: podcasts.apple.com/nz/podcast/the-best-ai-agents-are-secr…
🟢 Spotify: open.spotify.com/episode/5lT8hva20sVmJ0WiVCLAaO -
LangChain
@LangChain
What does production-ready infrastructure for agent execution actually require?
A secure agent computer needs to:
✅ Execute untrusted code safely
✅ Keep credentials and resources under control
✅ Give teams visibility into what ran and what it accessed
✅ Support fast iteration with reproducible environments
LangChain
@LangChain
On July 15, we’re hosting Build a Secure Computer for Your Agent, a technical webinar on sandboxed execution for agents.
We’ll walk through architecture patterns and use cases across coding agents, data analysis, and file processing.
Save your spot: events.langchain.com/register/build-a-secure-computer-for… -
LangChain reposted
Harrison Chase
@hwchase17
📕LLM Wiki Webinar with @BraceSproul @devstein64 @jeffreyhuber is now on YouTube!
Two of my favorite insights from this webinar:
"Wiki as a cache" - @devstein64: basically, the purpose of a wiki is to keep things that are commonly looked up or accessed more top of mind/readily available. What goes in the wiki should be updated and organized with that in mind
"Wiki is a set of hyperlinked pages" - @jeffreyhuber. The evolution of memory (in my view) has gone from:
1. single string
2. file
3. set of files in a directory (wiki)
as you think about scaling up - as files grow, the file structure matters less, and you need links between pages. The internet is a set of hyperlinked pages - so that definitely scales!
Watch it here: www.youtube.com/watch?v=Lsut4TCfygw -
LangChain
@LangChain
In 7 minutes, Partner Engineer Srimanth Tangedipalli shows you how to run Deep Agents Code inside a governed NemoClaw OpenShell Sandbox, with @NVIDIA Nemotron 3 Ultra served through @Baseten.
This media couldn't be displayed. See it on the original post.
LangChain
@LangChain
One command lets you set up Deep Agents Code as a governed blueprint on @NVIDIA NemoClaw, running Nemotron 3 Ultra.
Keep your source, your model, and your audit trail your own.
Everything you need to know.
www.langchain.com/blog/deep-agents-code-on-nemoclaw-a-gov… -
LangChain reposted
Sydney Runkle
@sydneyrunkle
see you in Boston!
LangChain
@LangChain
Cooking up 5 meetups and workshops in July! RSVP: luma.com/langchain
Find your city and bring a friend!
7/16 Berlin (with @Zalando and DB Engineering & Consulting)
7/21 NYC (with Clay)
7/21 London (with @sudip_kandel, @GirlsWhoML and @CopilotKit)
7/27 Boston
7/28 SF (with @Cognition)
7/30 Washington, D.C. (with @G2Xchange) -
LangChain
@LangChain
Don't miss this!
Andrew Wong
@41five
An agent that hides its work is just a confident guesser 👀
On July 14, @bromann (@LangChain) and I are showing how to build agents that actually show their work - live tool calls, reasoning steps, and trusted context from Box. No black boxes, no vibes-based answers.
Come see it live👇
events.box.com/box-webinars/box-langchain-2026 -
LangChain
@LangChain
One command lets you set up Deep Agents Code as a governed blueprint on @NVIDIA NemoClaw, running Nemotron 3 Ultra.
Keep your source, your model, and your audit trail your own.
Everything you need to know.
www.langchain.com/blog/deep-agents-code-on-nemoclaw-a-gov… -
LangChain
@LangChain
Cooking up 5 meetups and workshops in July! RSVP: luma.com/langchain
Find your city and bring a friend!
7/16 Berlin (with @Zalando and DB Engineering & Consulting)
7/21 NYC (with Clay)
7/21 London (with @sudip_kandel, @GirlsWhoML and @CopilotKit)
7/27 Boston
7/28 SF (with @Cognition)
7/30 Washington, D.C. (with @G2Xchange) -
LangChain
@LangChain
At Interrupt, @Rippling took the stage to share how they built and launched Rippling AI, a natural language assistant that can answer payroll, HR, and benefits questions across every system of record. youtu.be/3lb_4OEOykc
LangChain
@LangChain
📍 Interrupt NYC: September 24
📍 Interrupt London: October 13 interrupt.langchain.com -
LangChain reposted
NVIDIA Asia Pacific
@NVIDIAAP
Every company has specialized intelligence that sets it apart.
With @langchain, we’re helping enterprises turn that knowledge into specialized agents they can customize, secure and improve over time.
Learn more: nvda.ws/3SKngq1 -
LangChain
@LangChain
LangChain is hiring.
✅ Engineering
✅ GTM
✅ People + Finance
✅ Product
Check out our open roles + apply today ⤵️
www.langchain.com/careers -
LangChain
@LangChain
This agent drafts a cited VC investment memo in ~90 seconds for $0.40.
1️⃣ 4 LangGraph nodes powered by the @Perplexity_AI Agent API research financials, product, and market in parallel
2️⃣ A tool-less synthesizer writes the memo from only what they found
Deep dive: -
LangChain reposted
Brace
@BraceSproul
I just published a new video on OpenWiki Brains, specifically on the general-purpose brain mode.
In the video I dive into:
- configuring it locally
- its architecture
- what the docs look like
- new features like the 'open questions' agent file 👀
Check it out here: -
LangChain
@LangChain
.@NVIDIA built Nemotron to work inside agent harnesses.
We wanted to see how far we could take it.
An inside look on how we tuned Nemotron 3 Ultra inside Deep Agents. www.langchain.com/blog/tuning-the-harness-not-the-model-a… -
LangChain
@LangChain
LangSmith for Startups Spotlight: @finchlegal
✅ Powers pre-litigation for plaintiff personal injury firms, pairing experienced PI professionals with purpose-built agents.
✅ 10x growth over the past year, powering pre-litigation for leading plaintiff firms across the country.
✅ Uses LangSmith as their production observability layer — tracing, evaluating, and monitoring quality and cost across workflows like treatment check-ins, client communications, medical-record follow-up, document processing, and more.
Want to help build the future of consumer law? Finch is hiring exceptional engineers and operators in NYC: www.finchlegal.com/careers -
LangChain reposted
Mason Daugherty
@masondrxy
If you're in Boston, come hang out with us!
@sydneyrunkle and I are hosting two talks on what it takes to make agents reliable in production:
"The Art of Loop Engineering" — Sydney
"Why Open Models Have Crossed a Threshold" — yours truly
Harvard Square, July 27, 6 PM
RSVP: -
LangChain reposted
Sydney Runkle
@sydneyrunkle
calling all boston AI folks!
@masondrxy and i are talking about loop eng and open models, come with questions and ideas!
luma.com/iaie4h2x -
LangChain
@LangChain
Boston! You won’t want to miss this!
Sydney Runkle
@sydneyrunkle
calling all boston AI folks!
@masondrxy and i are talking about loop eng and open models, come with questions and ideas!
luma.com/iaie4h2x -
LangChain reposted
Viv
@Vtrivedy10
love what Brace is doing with open memory
- fully transparent how it’s generated, can use open models too
- integrates into your existing systems of work
- complements memory systems like in Claude and ChatGPT
- markdown first, take it wherever you want
- updates with you over time -
LangChain
@LangChain
"Whoever Jerry is, he was excellent."
That's a customer talking about an agent.
@PodiumHQ's Walker Ward sat down with our COO @j_schottenstein to share how LangGraph + LangSmith helped his team take their agents from prototype to production.This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
OpenWiki now supports general-purpose memory, in addition to code via the new "brains" mode. We believe both codebase and personal wikis are incredibly important for providing your agent memory, but have distinct workflows for configuring and running them.
Checkout our blog on the "brains" launch to see how we think about these two problems!
LangChain
@LangChain
And see the open-source OpenWiki repo here: github.com/langchain-ai/openwiki
-
LangChain reposted
Harrison Chase
@hwchase17
big update to openwiki - it now supports two modes:
- code brain: create a wiki for a code base
- personal brain: create a wiki for more general purpose tasks (email, web, etc)
webinar at 11am PST (5 minutes!) where we will be talking about wikis!
events.langchain.com/webinar/llm-wiki/ -
LangChain reposted
Colin Francis
@colifran_
happening now! go check it out
LangChain
@LangChain
Happening today!
events.langchain.com/webinar/llm-wiki/ -
LangChain
@LangChain
Read-only agents are easy to branch and test
Write access agents that touch production data are still an unsolved eval problem for many teams
From the latest Max Agency with @BenTannyhill
LangChain
@LangChain
Listen to the full conversation on your favorite podcasting platform.
⏯️ YouTube: youtu.be/YqjR4vQwbTc
🎧 Apple: podcasts.apple.com/nz/podcast/the-best-ai-agents-are-secr…
🟢 Spotify: open.spotify.com/episode/5lT8hva20sVmJ0WiVCLAaO -
LangChain reposted
-
LangChain reposted
Brace
@BraceSproul
OpenWiki Brains 0.1.0 is officially released!
We added a general-purpose memory brain to OpenWiki, in addition to the existing code brain. You can now use it to seamlessly setup a personal brain to track everything you do and are interested in.
OpenWiki brains are the easiest way to setup automatic memory agents for both codebases and general-purpose memory. Try it out today 👇 -
LangChain reposted
Harrison Chase
@hwchase17
this is going to be a fun one!
clay is one of the most ai-forward thinking companies i know of
lots to learn here if you're in NYC
LangChain
@LangChain
Join us for a LangChain + Clay meetup with @palashshah, @jeffbarg, Vyshu Khota, and Soroush Khadem. luma.com/jqif2hti
Palash will break down how he built a self-improving agent at LangChain, LangSmith Engine. He'll dive into how the agent turns production traces into durable memory updates that continuously improve an agent's behavior.
Then, the Clay team will give us a peek into how they built out their agentic eval harness, and more.
Interested in attending? RSVP while spaces last. -
LangChain
@LangChain
Join us for a LangChain + Clay meetup with @palashshah, @jeffbarg, Vyshu Khota, and Soroush Khadem. luma.com/jqif2hti
Palash will break down how he built a self-improving agent at LangChain, LangSmith Engine. He'll dive into how the agent turns production traces into durable memory updates that continuously improve an agent's behavior.
Then, the Clay team will give us a peek into how they built out their agentic eval harness, and more.
Interested in attending? RSVP while spaces last. -
LangChain
@LangChain
We built a plugin that traces every Claude Code session straight into LangSmith. Three commands, one JSON block, and every message, tool call, and subagent run shows up as an inspectable trace.
Setup takes about two minutes.
How are you currently debugging Claude Code when it goes off track?This media couldn't be displayed. See it on the original post.
LangChain
@LangChain
Also on YouTube: www.youtube.com/watch?v=jLOM_ahG78c
-
LangChain
@LangChain
Happening today!
events.langchain.com/webinar/llm-wiki/ -
LangChain reposted
Brace
@BraceSproul
We're hosting a meetup on agent memory and wikis July 28th at our SF office!
Come hear @jacobtpl and myself talk about the frontier research going on in these areas right now.
luma.com/mylwoabu -
LangChain reposted
NVIDIA
@nvidia
Every company has specialized intelligence that sets it apart.
With @langchain, we’re helping enterprises turn that knowledge into specialized agents they can customize, secure and improve over time.This media couldn't be displayed. See it on the original post. -
LangChain reposted
Neil Dahlke
@neildahlke
Everyone will focus on the benchmarks, but the real shift in the LangChain x NVIDIA blueprint is owning all three layers - the model, the harness, the runtime. When an agent breaks in prod you can open every layer instead of guessing against someone's black box.
LangChain
@LangChain
Introducing the NemoClaw Deep Agents Blueprint, a reference architecture for building open agent systems developed with @NVIDIA
✅ A fully open stack enterprises can own and customize
✅ Benchmark-leading performance
✅ Over 10x lower inference costs
Blog: www.langchain.com/blog/langchain-and-nvidia-launch-the-ne…
Video: youtu.be/Yy3JH6dDugcThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Viv
@Vtrivedy10
you can take open intelligence, open source primitives for infra+security, good evals, and run loops to make open models approach and exceed frontier level performance on tasks your company cares about at a fraction of the cost
and we think everyone should have the option to own this intelligence for themselves
LangChain
@LangChain
Introducing the NemoClaw Deep Agents Blueprint, a reference architecture for building open agent systems developed with @NVIDIA
✅ A fully open stack enterprises can own and customize
✅ Benchmark-leading performance
✅ Over 10x lower inference costs
Blog: www.langchain.com/blog/langchain-and-nvidia-launch-the-ne…
Video: youtu.be/Yy3JH6dDugcThis media couldn't be displayed. See it on the original post. -
LangChain reposted
NVIDIA AI
@NVIDIAAI
Excited to see this out in the open.
We've been working with @LangChain to tune their Deep Agents harness for Nemotron 3 Ultra.
The result is a fully open, customizable agent that delivers frontier performance at a fraction of the cost of leading closed models.
LangChain
@LangChain
We tuned the harness for @NVIDIAAI Nemotron 3 Ultra.
Benchmark-leading performance. 10x lower inference costs.
✅ An aggregate score of 0.86 at a cost of $4.48
✅ The closest-performing model: $43.48
www.langchain.com/blog/langchain-and-nvidia-launch-the-ne… -
LangChain
@LangChain
.@hwchase17 + Jensen Huang fireside chat.
A discussion on:
✅Today’s announcement
✅The blueprint
✅Open agent systems
✅The path to lower-cost enterprise agentsThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Brace
@BraceSproul
OpenWiki 0.0.3 just launched with tons of bug fixes and improvements!
Some notable changes:
- 'openai-compatible' provider to support any LLM provider that supports the OpenAI API spec! Thank you @bradhuffman
- deterministically skip updates if no code changes have been made. Shoutout @hudushy
- GitLab CI example workflow for running update jobs
- and many more smaller fixes and improvements!
lots more coming soon to OpenWiki. also very interested to hear what people want in a memory system -
LangChain
@LangChain
Excited to team up with partners across the AI infrastructure + enterprise ecosystem.
@EY_US
@baseten
@FireworksAI_HQ
@nebiusai
@CrusoeAI
@DeepInfra
@togethercompute
LangChain
@LangChain
Introducing the NemoClaw Deep Agents Blueprint, a reference architecture for building open agent systems developed with @NVIDIA
✅ A fully open stack enterprises can own and customize
✅ Benchmark-leading performance
✅ Over 10x lower inference costs
Blog: www.langchain.com/blog/langchain-and-nvidia-launch-the-ne…
Video: youtu.be/Yy3JH6dDugcThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Box
@Box
Box Agent uses @LangChain’s Deep Agents harness to bring specialized agents into the enterprise content platform.w
@NVIDIA + LangChain’s work with Nemotron 3 Ultra reinforces where AI is headed: open, tunable agent systems embedded where work happens.
Box makes them enterprise-ready.
Learn more about using Box and LangChain together in our upcoming webinar. events.box.com/box-webinars/box-langchain-2026
LangChain
@LangChain
Introducing the NemoClaw Deep Agents Blueprint, a reference architecture for building open agent systems developed with @NVIDIA
✅ A fully open stack enterprises can own and customize
✅ Benchmark-leading performance
✅ Over 10x lower inference costs
Blog: www.langchain.com/blog/langchain-and-nvidia-launch-the-ne…
Video: youtu.be/Yy3JH6dDugcThis media couldn't be displayed. See it on the original post. -
LangChain reposted
NVIDIA
@nvidia
Enterprises don’t just need agents that perform. They need agents they can shape, govern and improve as their business evolves.
That’s the idea behind our work with LangChain.
Read more: nvda.ws/4p6tDjg
LangChain
@LangChain
Introducing the NemoClaw Deep Agents Blueprint, a reference architecture for building open agent systems developed with @NVIDIA
✅ A fully open stack enterprises can own and customize
✅ Benchmark-leading performance
✅ Over 10x lower inference costs
Blog: www.langchain.com/blog/langchain-and-nvidia-launch-the-ne…
Video: youtu.be/Yy3JH6dDugcThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Inside the @NVIDIA NemoClaw Deep Agents Blueprint:
✅ Nemotron 3 Ultra: Open-weight model layer, customizable for your domain
✅ Deep Agents: Harness layer for planning, tool use, memory, + long-running tasks
✅ OpenShell Runtime: An open runtime you control, inspect, and govern -
LangChain
@LangChain
We tuned the harness for @NVIDIAAI Nemotron 3 Ultra.
Benchmark-leading performance. 10x lower inference costs.
✅ An aggregate score of 0.86 at a cost of $4.48
✅ The closest-performing model: $43.48
www.langchain.com/blog/langchain-and-nvidia-launch-the-ne… -
LangChain reposted
Jake Broekhuizen
@jakebroekhuizen
dcode <> nemotron 3 ultra <> OpenShell 🔥
Never been easier to own your agent IP!
LangChain
@LangChain
Introducing the NemoClaw Deep Agents Blueprint, a reference architecture for building open agent systems developed with @NVIDIA
✅ A fully open stack enterprises can own and customize
✅ Benchmark-leading performance
✅ Over 10x lower inference costs
Blog: www.langchain.com/blog/langchain-and-nvidia-launch-the-ne…
Video: youtu.be/Yy3JH6dDugcThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Harrison Chase
@hwchase17
Really excited to partner with @nvidia on the NemoClaw Deep Agents Blueprint
Deep Agents is a fully open source agent harness that we are tuning to make perform incredibly well with open models
LangChain
@LangChain
Introducing the NemoClaw Deep Agents Blueprint, a reference architecture for building open agent systems developed with @NVIDIA
✅ A fully open stack enterprises can own and customize
✅ Benchmark-leading performance
✅ Over 10x lower inference costs
Blog: www.langchain.com/blog/langchain-and-nvidia-launch-the-ne…
Video: youtu.be/Yy3JH6dDugcThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Alex Olsen
@its_ao
I cannot put into words how stoked I am on this
It's not an understatement to say that this blueprint could represent the future of enterprise inference
LangChain
@LangChain
Introducing the NemoClaw Deep Agents Blueprint, a reference architecture for building open agent systems developed with @NVIDIA
✅ A fully open stack enterprises can own and customize
✅ Benchmark-leading performance
✅ Over 10x lower inference costs
Blog: www.langchain.com/blog/langchain-and-nvidia-launch-the-ne…
Video: youtu.be/Yy3JH6dDugcThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Introducing the NemoClaw Deep Agents Blueprint, a reference architecture for building open agent systems developed with @NVIDIA
✅ A fully open stack enterprises can own and customize
✅ Benchmark-leading performance
✅ Over 10x lower inference costs
Blog: www.langchain.com/blog/langchain-and-nvidia-launch-the-ne…
Video: youtu.be/Yy3JH6dDugcThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
This Thursday!
@hwchase17 is moderating a discussion on LLM Wikis with @BraceSproul + @devstein64
Don't miss it!
RSVP @ events.langchain.com/webinar/llm-wiki/ -
LangChain reposted
Harrison Chase
@hwchase17
Good blog from viv on how improving agents (via rl, harness Eng, anything) boils down to a data mining problem over traces
-
LangChain reposted
Christian Bromann
@bromann
Very excited to talk with @41five about all things agent trust!
Andrew Wong
@41five
An agent that hides its work is just a confident guesser 👀
On July 14, @bromann (@LangChain) and I are showing how to build agents that actually show their work - live tool calls, reasoning steps, and trusted context from Box. No black boxes, no vibes-based answers.
Come see it live👇
events.box.com/box-webinars/box-langchain-2026 -
LangChain
@LangChain
.@SchneiderElec runs 60+ AI agents in production across 100+ countries, all traced through self-hosted LangSmith.
Their AI Assistant alone serves 160,000 employees.
How they built the LLMOps foundation to trust agents at that scale.
www.langchain.com/blog/how-schneider-electric-built-their… -
LangChain reposted
Harrison Chase
@hwchase17
deepagents is our newest open source project - an open source, model agnostic agent harness
this is maybe the most important academy course we've launched
LangChain
@LangChain
🎓 New course launch from LangChain Academy: Introduction to Deep Agents
✅ Learn what a harness is, and why agents need one
✅ Understand the 4 core capabilities of a harness
✅ Start building with Deep Agents
✅ Trace and deploy with LangSmithThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
🎓 New course launch from LangChain Academy: Introduction to Deep Agents
✅ Learn what a harness is, and why agents need one
✅ Understand the 4 core capabilities of a harness
✅ Start building with Deep Agents
✅ Trace and deploy with LangSmithThis media couldn't be displayed. See it on the original post.
LangChain
@LangChain
Enroll for free! academy.langchain.com/courses/foundation-introduction-to-…
-
LangChain
@LangChain
Agents are writing code, processing files, analyzing data, installing packages, and running multi-step workflows. To do that well, they need more than tool calls, they need a secure computer of their own.
But giving agents an execution environment introduces new questions around security, isolation, observability, and control.
Join us on July 15 for Build a Secure Computer for Your Agent, a technical webinar on how to safely give agents the workspace they need to do real work.
Register here: events.langchain.com/register/build-a-secure-computer-for… -
LangChain reposted
Brace
@BraceSproul
Your traces matter a lot more than you might think. Great article on this from Mr. Agents himself @Vtrivedy10
-
LangChain reposted
LangChain
@LangChain
We're taking Interrupt on the road this fall with two new stops in NYC and London.
🗽 NYC: September 24
🇬🇧 London: October 13
Get your tickets now! interrupt.langchain.com -
LangChain
@LangChain
ICYMI: At @aiDotEngineer World’s Fair, @vtrivedy10 took the stage to share why data mining from traces is one of the highest leverage muscles companies can build to understand their agents, curate data at scale, & then run improvement loops.
-
LangChain reposted
Harrison Chase
@hwchase17
devin has a hot take that wikis are NOT the right abstraction
should be a fun webinar!!
Devin Stein
@devstein64
does your agent need a wiki? should humans and agent use the same wiki? should you have one wiki or many wikis?
if you're wiki-curious, join @BraceSproul @hwchase17 and I on Thursday!
you don't want to miss it @dosu_ai 🤝 @LangChain -
LangChain
@LangChain
When your agents take a tumble, LangSmith helps them get back up.
LangSmith Evaluation lets you evaluate performance with real production data so you can understand every spill and improve agent quality.
LangChain
@LangChain
More on LangSmith Evaluation www.langchain.com/langsmith/evaluation
-
LangChain
@LangChain
We're taking Interrupt on the road this fall with two new stops in NYC and London.
🗽 NYC: September 24
🇬🇧 London: October 13
Get your tickets now! interrupt.langchain.com -
LangChain
@LangChain
Last minute webinar this Thursday
Will cover all things related to "LLM Wikis"
Harrison Chase
@hwchase17
🚨Emergency webinar: LLM Wikis and how to give your agent memory
LLM Wikis are so hot right now - OpenWiki by @BraceSproul up to nearly 7k GitHub stars in less than a week
I'll be chatting with Brace and @devstein64 about Wikis this Thursday: events.langchain.com/webinar/llm-wiki/ -
LangChain reposted
Harrison Chase
@hwchase17
🚨Emergency webinar: LLM Wikis and how to give your agent memory
LLM Wikis are so hot right now - OpenWiki by @BraceSproul up to nearly 7k GitHub stars in less than a week
I'll be chatting with Brace and @devstein64 about Wikis this Thursday: events.langchain.com/webinar/llm-wiki/ -
LangChain
@LangChain
How to run agent evals with @harborframework and LangSmith sandboxes, full traces included.
This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Product Manager @BenTannyhill on why LangSmith Engine routes trace investigation through screener + verifier sub agents instead of letting the main agent read everything.
LangChain
@LangChain
Listen to the full conversation on your favorite podcasting platform.
⏯️ YouTube: youtu.be/YqjR4vQwbTc
🎧 Apple: podcasts.apple.com/nz/podcast/the-best-ai-agents-are-secr…
🟢 Spotify: open.spotify.com/episode/5lT8hva20sVmJ0WiVCLAaO -
LangChain reposted
Brace
@BraceSproul
OpenWiki is at 1.7k stars in just 3 days!
Right now it's just for codebases, but we're working to expand it to everything for memory.
What do you want to see in a general purpose memory wiki agent?
github.com/langchain-ai/openwiki -
LangChain reposted
Viv
@Vtrivedy10
my fave question, talked about this coding agent Eval+Improvement loop infra + UX in my AIE talk yesterday!
biased but LangSmith is the best spot to Eval + continuously improve your coding agents, and we want to make it better so would love any feedback :)
we eval all of our coding agents there --> supports Codex, Claude Code, OpenCode, Deep Agents, Pi, etc all into Tracing, sandbox infra for running evals, metrics + datasets for storing everything, and
imo the hardest parts of doing coding agent evals are:
1. having infra to easily store, update, run, and share coding agent evals
2. building a clear understanding of agent behavior & failure modes across all of the rollouts. you can use a coding agent + the langsmith-cli to look through all the data
or if teams want a managed, they can use LangSmith Engine to read every single trace from every eval, see what went wrong, prepare a report for you, and propose new changes directly in the code to fix issues
3. build new and better evals for your coding tasks. there's many ways to do this, but we find that looking through existing failure modes from eval runs and prod is a really good grounded way to measure where agents lack today and turn that data into
we recently launched a LangSmith x @harborframework integration to double down on making it super easy for LangSmith users to improve their coding agents over time with built-in infra for running large-scale containerized evals so you can fully reproduce your tasks and read all of the traces
www.langchain.com/blog/unified-stack-for-evaluating-agents
i think most evals in the future will be shaped as environments and letting agents do real work in them, because agents are doing way more complex things and we need the eval shape to mirror how they work with us
if there's any experience you're looking for would love to chat
we have a ton of work underway on making each part of coding agent evaluation + improvement easier over time (as teams eval their coding agents for months and years)
as an aside --> evals are literally the training data for agents. the behaviors that we measure and reward in the evals directly get transferred as model/agent behavior as we hill-climb them. so i think there's nothing more important than making it easy to build good evals over time :)
Michael Thiessen
@MichaelThiessen
@Vtrivedy10 do you know of any eval platforms that work with coding agents?
Unless I'm blind, everything looks like it's product-agent focused.
I need something that will work with coding agents on complex R&D tasks.
(currently building my own so we can properly eval our harness) -
LangChain
@LangChain
Setup OpenWiki once, then forget about it. It'll automatically update its docs as you make changes to your codebase
OpenWiki is the easiest way to increase your coding agent performance, and it's fully OSS:
- open harness (deepagents)
- open CLI/workflows (OpenWiki repo)
- runs on open models (GLM 5.2, Kimi K2.7, etc)
Checkout the OSS repo 👇
Brace
@BraceSproul
OpenWiki is designed to run in the background, without you needing to think about it
It'll generate docs, update your AGENTS.md so your agent automatically knows how to read the docs, and update itself automatically. All of this allows you to set it up once, then forget about it!
To let it update itself automatically, we've included a GitHub workflow you can copy into your repo which will run OpenWiki nightly to update its docs!
Check it out today and start using OpenWiki locally: github.com/langchain-ai/openwiki/blob/main/examples/openw…
LangChain
@LangChain
GitHub repo: github.com/langchain-ai/openwiki
-
LangChain reposted
Brace
@BraceSproul
OpenWiki is designed to run in the background, without you needing to think about it
It'll generate docs, update your AGENTS.md so your agent automatically knows how to read the docs, and update itself automatically. All of this allows you to set it up once, then forget about it!
To let it update itself automatically, we've included a GitHub workflow you can copy into your repo which will run OpenWiki nightly to update its docs!
Check it out today and start using OpenWiki locally: github.com/langchain-ai/openwiki/blob/main/examples/openw… -
LangChain
@LangChain
Your coding agent bill doubled and nobody can tell you why.
Here's the actual reason: Claude Code, Cursor, and Copilot all log activity in different formats. The second your team uses more than one (they will), your visibility goes to zero.
We built LangSmith to fix this. It traces Claude Code, Codex, Cursor, Copilot, Pi, and OpenCode into one consistent format.
✅ Same trace tree
✅ Same metadata
✅ Same query syntax
…regardless of which tool actually ran the session.
Everything you need to know: www.langchain.com/blog/fix-your-coding-agent-bill -
LangChain
@LangChain
🎧A look at how we built LangSmith Engine with @hwchase17 + @bentannyhill
Harrison Chase
@hwchase17
Special episode with @bentannyhill on the Max Agency podcast. A month ago, his team shipped LangSmith Engine, our agent that hunts through your agent's failures, prioritizes issues, and drafts the fix.
We dive into the architecture decisions -- from how we used sandboxes to how we created subagents. And we also discuss an interesting challenge -- how we were able to build evals for an agent that never stops running.
Check out the full conversation ⤵️
⏯️ YouTube: youtu.be/YqjR4vQwbTc
🎧 Apple: podcasts.apple.com/nz/podcast/the-best-ai-agents-are-secr…
🟢 Spotify: open.spotify.com/episode/5lT8hva20sVmJ0WiVCLAaOThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Jacob Lee
@Hacubu
Some of LangSmith's biggest and most engaged customers trace with:
- @aisdk
- Claude Agents SDK
- OpenAI Agent SDK
- No framework at all
And I'm sure we'll see tons of pi.dev in the near future
We provide a first-class experience no matter how you build!
Julia Schottenstein
@j_schottenstein
the most common FUD our competitors throw against LangSmith is that it only works if you use our open source.
but actually as of this month, a MAJORITY of our active self-serve customers build without a LangChain open source framework / harness.
LangSmith is model, framework, and cloud neutral, and we have the numbers to show it! -
LangChain
@LangChain
If you want a multi-agent system so one team can own one agent, you're shipping your org chart instead of building the best possible product.
@zackrw on why @SierraPlatform defaults to one agent per brand.
LangChain
@LangChain
Listen to Max Agency with @hwchase17 on your favorite podcasting platform:
🎧 Apple: podcasts.apple.com/nz/podcast/the-best-ai-agents-are-simp…
🎧 Spotify: open.spotify.com/episode/2jWGobitRmBQUygBiVhD2c
▶️ YouTube: youtu.be/uCKhOmth2ms -
LangChain reposted
Caspar
@caspar_br
Imagine if, as an engineer, your working memory of a codebase was wiped every time you started a new ticket
Sounds ridiculous, but this is functionally what happens to agents whenever we start a new thread
Maybe your company is ahead of the curve and has committed context like 𝙰𝙶𝙴𝙽𝚃𝚂.𝚖𝚍 and skills. But most problems are both complex and niche, so agents still waste a ton of tokens rebuilding the context they need to solve them
Code is the current state of a system, but it is not enough to understand the whole thing. OpenWiki adds the missing context layer and maintains it automatically, so agents know how pieces fit together, which patterns your company values, and where to look before making changes. Really excited for this release -
LangChain reposted
Viv
@Vtrivedy10
come by the session to yap about Continual Learning, Auto-Research and why/how you gotta “look at the data” to do any of it 👀
LangChain
@LangChain
At @aiDotEngineer World's Fair? Don't miss @Vtrivedy10's session at 1:55pm!
More info: www.ai.engineer/worldsfair/schedule?q=langchain&session=a… -
LangChain
@LangChain
At @aiDotEngineer World's Fair? Don't miss @Vtrivedy10's session at 1:55pm!
More info: www.ai.engineer/worldsfair/schedule?q=langchain&session=a… -
LangChain
@LangChain
Interrupt 2026 is coming to New York City and London this fall.
→ Hear from engineers + teams shipping agents in production
→ Keynotes from @hwchase17 + industry leaders on what's next for agents
→ Hands-on workshops with LangChain product experts
→ First look at new LangChain products
Tickets go on sale next week and seats are limited.
Sign up to get notified when tickets drop:
interrupt.langchain.com -
LangChain
@LangChain
LangSmith Sandboxes: Now available on AWS SaaS.
Available across all LangSmith environments including APAC and AWS
Give your agents a secure computer today:
docs.langchain.com/langsmith/sandboxes -
LangChain
@LangChain
At @aiDotEngineer World's Fair? @jakebroekhuizen is speaking at 11:40am!
More info: www.ai.engineer/worldsfair/schedule?q=langchain&session=a… -
LangChain
@LangChain
OpenWiki is the easiest way to document your codebase, built specifically for agents to consume.
OpenWiki can generate repo docs, automatically update them as your codebase evolves, and do Q&A over your docs and codebase!
github.com/langchain-ai/openwiki -
LangChain reposted
Nick Hollon
@nickhollon10
this video goes into more detail on how our integrations with @harborframework actually work! if the blog yesterday caught your eye definitely watch this! thanks for all the help @kobe0938 @alexgshaw!!
LangChain
@LangChain
.@harborframework is now wired directly into Deep Agents, LangSmith Sandboxes, and LangSmith Observability.
This video contains everything you need to know about this integration in 10 mins:
✅ What is Harbor?
✅ How to integrate a Deep Agent with Harbor
✅ How to view results in LangSmithThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Brace
@BraceSproul
Just published a YouTube video walking through OpenWiki. Setup and configuration, how it works, what it targets, and how to automatically keep your docs up to date.
Check it out, and see the OSS GitHub repo with the OpenWiki code!
youtu.be/nIVu3zfYprI -
LangChain reposted
-
LangChain reposted
Brace
@BraceSproul
Excited to release OpenWiki!
OpenWiki is the easiest way to generate and maintain documentation in your codebase, built specifically for agents to consume
Get started with a single command: `openwiki --init` -
LangChain
@LangChain
.@harborframework is now wired directly into Deep Agents, LangSmith Sandboxes, and LangSmith Observability.
This video contains everything you need to know about this integration in 10 mins:
✅ What is Harbor?
✅ How to integrate a Deep Agent with Harbor
✅ How to view results in LangSmithThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
You can now run recursive language model (RLM) workflows in Deep Agents.
Everything you need to know in 6 minutes from @sydneyrunkle.This media couldn't be displayed. See it on the original post.
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Alex Olsen
@its_ao
Open models are the future, and that future starts now
GLM 5.2 has become my daily driver. here’s how and why:
LangChain
@LangChain
You can start building with GLM 5.2 in minutes:
1️⃣ Download dcode
2️⃣ Select your model (GLM 5.2)
3️⃣ Enter your API key
…and you're ready to go with frontier performance on open weights.
Great demo from @its_aoThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Mason Daugherty
@masondrxy
Team is on fire this week
Another banger for your reading list!
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Sydney Runkle
@sydneyrunkle
you can now build RLMs with Deep Agents!
the more context agents accumulate, the worse they perform, a phenomenon called context rot.
RLMs, proposed by @a1zhang from MIT, help: instead of working model turn by turn, the model runs code in a REPL that dynamically dispatches subagents and recurses over the input.
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
You can start building with GLM 5.2 in minutes:
1️⃣ Download dcode
2️⃣ Select your model (GLM 5.2)
3️⃣ Enter your API key
…and you're ready to go with frontier performance on open weights.
Great demo from @its_aoThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
With LangSmith, @Pendoio catches 60% of agent failures before customers ever see them.
Every morning, their team
📊 Opens the trace dashboard
✅ Finds the gap between customer needs + what Novus supports
💻 Builds new eval sets
How they’re tracing Novus from user behavior to code fixes:This media couldn't be displayed. See it on the original post. -
LangChain reposted
LangChain
@LangChain
Building voice agents can come with tradeoffs.
💬 Better convos w/ speech to speech models
Vs
🥪 More reliable harnesses w/ sandwich architecture
How to build a voice research agent with both:
✅ Gemini Live: low-latency, natural-sounding convos
✅ Deep Agents: long-running research tasks
✅ LangSmith: full tracing & observabilityThis media couldn't be displayed. See it on the original post. -
LangChain reposted
Madhur Prashant
@MadhurPras81001
At Antimetal, we are building the autonomous system for production engineering.
We build a unified world model of your environments and run a suite of agents to improve your system over time.
If you are interested and want to learn more, do reach out!
LangChain
@LangChain
LangSmith for Startups Spotlight: @Antimetal
Antimetal is an autonomous system to manage your software in production. It builds a world model of your environment -- then its agents continuously monitor to prevent issues, triage alerts, and ship fixes.
They are currently in private beta and used by engineering, infrastructure, and SRE teams at startups and the Fortune 500.
They build with Deep Agents and leverage LangSmith for tracing, evaluations, and sandboxes.
Speak to their engineering team for a demo at antimetal.com/demo. -
LangChain
@LangChain
LangSmith for Startups Spotlight: @Antimetal
Antimetal is an autonomous system to manage your software in production. It builds a world model of your environment -- then its agents continuously monitor to prevent issues, triage alerts, and ship fixes.
They are currently in private beta and used by engineering, infrastructure, and SRE teams at startups and the Fortune 500.
They build with Deep Agents and leverage LangSmith for tracing, evaluations, and sandboxes.
Speak to their engineering team for a demo at antimetal.com/demo. -
LangChain
@LangChain
Building voice agents can come with tradeoffs.
💬 Better convos w/ speech to speech models
Vs
🥪 More reliable harnesses w/ sandwich architecture
How to build a voice research agent with both:
✅ Gemini Live: low-latency, natural-sounding convos
✅ Deep Agents: long-running research tasks
✅ LangSmith: full tracing & observabilityThis media couldn't be displayed. See it on the original post.
LangChain
@LangChain
Watch the full video on YouTube: www.youtube.com/watch?v=aV0ctbnDUCM
-
LangChain reposted
Sydney Runkle
@sydneyrunkle
agents that can write code can solve problems more reliably
but you need to make sure you execute that untrusted code in a safe environment
here’s how we enable that w a lightweight code interpreter!
Hunter Lovell
@huntlovell
-
LangChain reposted
Viv
@Vtrivedy10
a very cool Harbor x LangSmith flow I love to help you “look at the data”:
1. you do evals or rollouts for RL
2. all reward metrics and traces and rollouts get automatically propulates into Experiments & Tracing projects
3. we send agents (or use Engine) to understand how training or harness engineering affects agent behavior over each iteration
Evals ~= Environments! more traditional Evals will become environment shaped as teams measure their complex agents rigorously
evaluating agents in a way that reflects prod means creating an environment that mirrors prod
late last year, we went all in on Harbor for our open source evals! and now anyone can also do the same and get all of that neatly organized to view across your team in LangSmith
for both our Eval runs and RL rollouts, it’s very cool to have a shared source of truth where the team can understand how each targeted experiment affects our agent -
LangChain
@LangChain
At @aiDotEngineer World's Fair?
We're hosting a World Cup Watch Party at The Harlequin today from 11-7 with @baseten and @reductoai
One block from the venue!
LangChain
@LangChain
⚽️ More info + RSVP: luma.com/aie-worldcup
-
LangChain reposted
Hunter Lovell
@huntlovell
we launched code interpreters for deep agents last month. Basic idea is to let agents plan, delegate, and organize context using code instead of chained tool calls
Code interpreters don't need a sandbox, but we still need a way to securely run that code! (and running untrusted code is a famously hard problem)
Here's the writeup on how we're looking to do just that:
Hunter Lovell
@huntlovell
-
LangChain
@LangChain
Giving agents the ability to write code makes them dramatically more capable.
It also makes security a lot harder.
At LangChain, we have spent a lot of time this year figuring out how to do both. www.langchain.com/blog/running-untrusted-agent-code-witho… -
LangChain reposted
Brace
@BraceSproul
Memory has somehow consistently been the most exciting area of agent development over the last 3 years (imo), and it's still a largely unsolved problem!!
Wiki's are the biggest advancement I've seen in a long time, and I'm super excited to push more on research here. Great article on them from Harrison, and stay tuned for more memory + wiki OSS from LangChain in the coming days!
Harrison Chase
@hwchase17
-
LangChain
@LangChain
.@harborframework can now integrate directly with Deep Agents, LangSmith Sandboxes, and LangSmith Observability.
You need to run agents in a real, reproducible, isolated environment, many times in parallel, with a deterministic check at the end.
Harbor solves this problem.
Everything you need to know⤵️ -
LangChain reposted
Harrison Chase
@hwchase17
-
LangChain reposted
Harrison Chase
@hwchase17
harbor is a great framework for running evals for long running, stateful agents
its becoming industry standard, powering benchmarks like terminal bench 2
we've integrated deeply with harbor: across LangSmith, Sandboxes, and Deep Agents -
LangChain reposted
Nick Hollon
@nickhollon10
joined @LangChain almost three months ago and have spent a ton of time thinking about evals! i think that harbor is the future of eval frameworks so really enjoyed being able to bring it into tighter integration with our stack!!
-
LangChain
@LangChain
At Interrupt, @LATAMAirlines took the stage to share how their B2C travel agent, built on LangGraph, handles thousands of daily users.
An essential watch for anyone interested in building agents in the aviation industry. youtu.be/RnLCl3ilRgo
LangChain
@LangChain
Want to watch more sessions from Interrupt? Access our archives today. interrupt.langchain.com/recordings
-
LangChain reposted
LangChain
@LangChain
Turn any LangGraph agent into a voice agent, no rebuilding required.
A how-to on:
✅ Making LangGraph the LLM layer of a voice agent built with @pipecat_ai
✅ Tracing your entire voice agents – from audio streams to end-of-speech detection to tool calls – in LangSmith
⏯️ -
LangChain reposted
Neil Dahlke
@neildahlke
The thing that breaks production agents usually isn't reasoning. It's volume. Hundreds of docs or thousands of datapoints is where turn-by-turn tool calling falls apart.
Dynamic subagents in Deep Agents fix it by letting the agent orchestrate in code, writeup by @sydneyrunkle
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Viv
@Vtrivedy10
auto-research at scale, agent-scale data mining, voting via mixture of agents
doing these effectively means your Harness should support dispatching & orchestrating many subagents on massive scale data/tasks
“read all 100k traces & experiment logs. find ways to reduce our token spend by 50% and while maintaining accuracy”
doing this reliably means agents need to dispatch other agents programmatically
we want every team to have this ability + find it useful for our data heavy (and other) tasks that Sydney lays out - so it’s available in Deep Agents today!
give it a spin —> I’ll say to do this well requires more up front planning. focusing more on how problems should be decomposed and how they should be verified
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Adam Łucek
@AdamRLucek
Actually a critical concept to understand for anyone seriously building agents. I work on one that can spawn up to 50 subagents at once- the only way to do that reliably and effectively is with this feature
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
A useful ad hoc task in LangSmith Fleet can graduate into a Specialized Agent without starting from a blank config screen.
Ask it to handle something in General Purpose Chat. It works. The work keeps coming back. Fleet helps you turn it into a Specialized Agent with the right instructions, tools, triggers, and memory already in place.
www.langchain.com/blog/why-fleet-has-both-general-purpose… -
LangChain
@LangChain
At @aiDotEngineer World's Fair?
Come find us at Booth U-G19!
If you're working on agents in production, thinking through your eval setup, or just want to see what we've been shipping, we'd love to connect. -
LangChain reposted
Ben Levine
@blevined
Psyched to see this work out in the world! We’re building state-aware agent harnesses, turning agent evals from a post-hoc grade into real-time control signals to improve financial outcomes for everyone.
gif by me and @claudeai
LangChain
@LangChain
Most agents are graded at the end. Pass/fail, resolved/abandoned.
@GetCandidly built something different: a model that reads the conversation while it's happening and steers toward resolution in real time.
Full story: www.langchain.com/blog/how-candidly-built-state-aware-age… -
LangChain
@LangChain
Most agents are graded at the end. Pass/fail, resolved/abandoned.
@GetCandidly built something different: a model that reads the conversation while it's happening and steers toward resolution in real time.
Full story: www.langchain.com/blog/how-candidly-built-state-aware-age… -
LangChain reposted
Christian Bromann
@bromann
I've experimented with this quite a bit and have been impressed by how this has unlocked a lot of scenarios for my agents to work more effectively. 🤯
You can now bring a skill to the orchestrator to suggest how tasks should be structured and approached effectively 😍
Sydney Runkle
@sydneyrunkle
-
LangChain reposted
Mason Daugherty
@masondrxy
Masterclass deep-dive into programmatic spawning of subagents - highly encourage taking a look! @colifran_ dropping elite ball knowledge per usual
Sydney Runkle
@sydneyrunkle
this blog covers 6 common workflow patterns we see with dynamic subagents (fanout and synthesize, adversarial verification, etc)
if you'd prefer to learn w/ a video, check out this excellent guide from @colifran_ with examples and trace walk throughs!This media couldn't be displayed. See it on the original post.
Sydney Runkle
@sydneyrunkle
-
LangChain
@LangChain
Deep Agents now supports dynamic subagents. Instead of invoking subagents with tool calls, the main agent writes orchestration code to coordinate work at scale.
This enables workflows like processing hundreds of documents or analyzing thousands of datapoints with deterministic coverage.
6 common orchestration patterns and how to get started today ⤵️
Sydney Runkle
@sydneyrunkle
LangChain
@LangChain
Want to learn more? Here's a great demo from @colifran_
www.youtube.com/watch?v=5AkdMangfNk -
LangChain
@LangChain
Turn any LangGraph agent into a voice agent, no rebuilding required.
A how-to on:
✅ Making LangGraph the LLM layer of a voice agent built with @pipecat_ai
✅ Tracing your entire voice agents – from audio streams to end-of-speech detection to tool calls – in LangSmith
⏯️ -
LangChain
@LangChain
At Interrupt, @Box Co-Founder & CEO @Levie sat down with @hwchase17 to dig into why AI agents are thriving in software development, but still struggling to break through in the rest of the enterprise. youtu.be/agSRMrhNTf4
This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
Selling a membership or closing a car sale? Outcome-based pricing makes sense. The value is real and measurable.
Answering a knowledge lookup? Usage-based is simpler and probably right.
@zackrw from @SierraPlatform on how AI pricing should actually map to value delivered.
LangChain
@LangChain
Listen to Max Agency with @hwchase17 on your favorite podcasting platform:
🎧 Apple: podcasts.apple.com/nz/podcast/the-best-ai-agents-are-simp…
🎧 Spotify: open.spotify.com/episode/2jWGobitRmBQUygBiVhD2c
▶️ YouTube: youtu.be/uCKhOmth2ms -
LangChain
@LangChain
.@benchling Head of AI @nlarusstone on why understanding LLMs is closer to biology than software engineering.
-
LangChain
@LangChain
In the latest Max Agency, @hwchase17 asked @benchling Head of AI @nlarusstone about the biggest blockers in building agents for scientific work.
This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
.@Box Agent is built on Deep Agents.
✅ Cross-library search
✅ Multi-doc synthesis
✅ Structured reports
✅ All within Box's existing security and permissions model -
LangChain
@LangChain
Head of AI @nlarusstone on the patterns @benchling uses to look at production traces.
-
LangChain
@LangChain
.@nlarusstone from @benchling on how they cross-check answers between models to get more out of each one.
LangChain
@LangChain
Watch the latest Max Agency with @hwchase17:
⏯️ YouTube: www.youtube.com/watch?v=RjpTrffSMjE
🎧 Apple Podcasts: podcasts.apple.com/us/podcast/the-tool-design-tricks-behi…
🎧 Spotify: open.spotify.com/episode/2bFEj2W290bk2JW1zC6wyp -
LangChain
@LangChain
On the latest episode of Max Agency, @hwchase17
sat down with @nlarusstone, Head of AI at @benchling for a conversation on building agents for scientific work.
⏯️ YouTube: www.youtube.com/watch?v=RjpTrffSMjE
🎧 Apple Podcasts: podcasts.apple.com/us/podcast/the-tool-design-tricks-behi…
🎧 Spotify: open.spotify.com/episode/2bFEj2W290bk2JW1zC6wypThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
This macroeconomic research agent powered by Deep Agents, LangSmith, and the @youdotcom Finance Research API:
✅ Analyzes GDP data
✅ Detects anomalies
✅ Investigates structural & cyclical drivers at the sector level
✅ Produces structured, cited briefings -
LangChain
@LangChain
Excited to have @Box in the Interrupt lineup!
CEO Aaron @Levie will take the stage with @hwchase17 to discuss agents in the enterprise. If you’re attending our agent conference, don’t miss their session on May 13th! -
LangChain
@LangChain
.@Redisinc is returning to Interrupt!
Say hello to their team at the expo hall to learn about their fast memory layer for chatbots and AI agents as well as their ready-to-use tools for building AI apps. -
LangChain
@LangChain
Build the future of Agentic AI with LangChain & @MongoDB at our hackathon in London.
Join us at CodeNode on Saturday, 2nd May for our Agentic Evolution hackathon.
- Prizes: Compete for a share of £15,000 in cash and partner credits.
- Main Stage: The top teams will pitch their ideas on the main stage at MongoDB.local London in front of a VIP judging panel
Move beyond the chat box and build something autonomous.
Register 👉 cerebralvalley.ai/e/mongo-db-london-hackathon -
LangChain
@LangChain
LangSmith for Startups gives early stage teams the tooling and community they need to iterate faster and win bigger.
Observe, evaluate, and deploy your agents — now with free credits to help you get started.
Our Scale program offers:
- $10,000 in LangSmith credits
- Exclusive programming: Technical sessions directly with the LangChain team
- Community: High-level networking with peer founders - Visibility: Product feature & recruiting spotlights on our social platforms
We’re proud to partner with the world’s leading investors to bring this to life: @sequoia, @ycombinator, @AmplifyPartners, @spc, @MayfieldFund, @NEA, @BessemerVP, @firstround, @BasisSet, @Lux_Capital, @kleinerperkins, @AforeVC, @Accel, @khoslaventures, @AudaciousHQ, @benchmark and @conviction.
How to join: If you're a startup backed by one of our partner VCs and Series A or earlier stage, you qualify! Join today 👉 langchain.com/startups -
LangChain
@LangChain
🔥 Hear Harrison's hot takes in this latest podcast with VentureBeat 🔥
- LangChain's origin story: from `langchain` to agent engineering platform serving half of the Fortune 10
- Why OpenClaw works where AutoGPT failed
- LangChain Deep Agents unpacked: planning, subagents, file systems, prompting
- Skills vs. tools vs. subagents
- Context engineering: what the LLM sees vs what developers see
- File systems for context management vs. AutoGPT's approach
YouTube ➡️ www.youtube.com/watch?v=53gPwkcIsXQ&list=PLMQoSwszBxm5dCv…
Spotify ➡️ open.spotify.com/show/4Zti73yb4hmiTNa7pEYls4 -
LangChain
@LangChain
Yesterday we launched the Agent Builder Template Library.
Agent templates are more than a starting point. They’re ready to deploy and fully customizable by giving your agent feedback in plain language.
🧵 for highlights
Try Agent Builder: www.langchain.com/langsmith/agent-builder -
LangChain
@LangChain
🚀 LangSmith Agent Builder is GA – and we’re moving fast
Agent Builder lets you build an agent with a simple prompt. But sometimes you want something that’s ready to go.
Today we’re introducing the Agent Builder Template Library: ready-to-deploy agents built with the companies who know their domains best.
Templates include:
Tavily: competitor intelligence
PagerDuty: on-call triage
Box: document intake review
+more
Supports: Gmail, Google Calendar, Slack, Linear, Pylon, GitHub, LinkedIn, X, Tavily, Exa + any app with an MCP server
👉 Try Agent Builder: www.langchain.com/langsmith/agent-builder
📣 Read the announcement: blog.langchain.com/introducing-agent-builder-template-lib…This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
LangSmith Agent Builder is generally available 🎉
It’s surprisingly easy to build agents now. Even a VC can do it…👇
Try it free: smith.langchain.com/agents?skipOnboarding=true
Read the announcement: blog.langchain.com/langsmith-agent-builder-generally-avai…This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
📊 Learn how to observe & evaluate agents on LangChain Academy 📊
Testing applications is essential to the development lifecycle, but LLM systems are non-deterministic – you can’t always predict how they will behave.
Add multi-turn interactions and tool-calling agents, and testing agents becomes even more complex than traditional software testing 😵💫
To address this challenge, LangSmith provides a comprehensive platform for agent engineering that helps teams use live production data for continuous testing and improvement 🚀
In our quickstart course, LangSmith Essentials, you’ll learn how to observe, evaluate, and deploy your agent in less than 30 minutes.
Enroll for free ➡️ academy.langchain.com/courses/quickstart-langsmith-essent…This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
How @cursor_ai's engineering team builds and ships 🚀
Harrison and Jason Ginsberg, Cursor’s Head of Engineering, sit down for a fireside chat on how Cursor builds and ships the next generation of AI coding agents.
Learn about:
- How Cursor's team ships features in 2-3 days: the hiring bar, process decisions, and dogfooding culture that makes it possible
- Why agents that evaluate themselves (like debug mode) change the game—and how they think about agents looping until they get it right
- Their take on the "right way to use Cursor": why short prompts and killing/restarting beats elaborate multi-stage planning
Hear the full chat 👉 youtu.be/dKSGK-fPFyUThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
We surveyed 1,300+ developers and business leaders to understand how AI agents are actually being built, tested, and shipped to real users.
What are the most popular agent use cases? Biggest challenges when building agents? And who's really using observability or evals tooling?
Read the full report here ➡️ www.langchain.com/state-of-agent-engineering/
Here's 6 key insights below 🧵👇 -
LangChain
@LangChain
💻Sandboxes for DeepAgents
We're excited to launch Sandboxes for DeepAgents, a new set of integrations that allow you to safely execute arbitrary DeepAgent code and bash commands in remote sandboxes.
Supports @RunloopAI @daytonaio @modal
Your DeepAgent runs locally (or wherever you want), but when it needs to execute code, create files, or run commands, those operations happen in the remote sandbox.
Blog: blog.langchain.com/execute-code-with-sandboxes-for-deepag…
Docs: github.com/langchain-ai/deepagents
Video: youtu.be/CejntUP3muU -
LangChain
@LangChain
Why we built LangSmith for improving agent quality
As more agents move into production, teams need to move beyond vibe-checking and bring rigor to how they understand agent behavior at scale.
In this video, the LangSmith engineering team and Harrison (@hwchase17) sit down to discuss how production observability for agents is evolving to meet this challenge. They cover:
• How Insights Agent automatically discovers usage patterns in your production traces
• Measuring user sentiment and tool call trajectories for end-to-end interactions
• Why both offline and online evals matter as agents move into critical use cases
Watch the full conversation here: youtu.be/J5X6HowopEc/ -
LangChain
@LangChain
🔥 Build advanced AI agents - LangGraph & Cerebras tutorial for Beginners 🔥
Check out our new YouTube tutorial on freeCodeCamp, created in partnership with @cerebras.
In this tutorial, our own @RLanceMartin walks you through building multi-agent workflows using LangGraph and Cerebras. This tutorial is great for beginners who want to level up quickly. -
LangChain
@LangChain
Learn Context Engineering Best Practices with LangChain & @manusai 🛠️
We’re excited to share that @RLanceMartin (Founding Engineer, LangChain) and @peakji (Co-founder & Chief Scientist, Manus and recent MIT Innovator Under 35) are teaming up for an upcoming webinar on context engineering best practices.
In this session, you’ll learn how to:
- Manage context windows effectively
- Optimize performance for complex workflows
- Build reliable, scalable agents
Hear practical insights from Yichao on context engineering, and hear from Lance how LangGraph can help you with context engineering challenges.
RSVP here → luma.com/819i5ime -
LangChain
@LangChain
Introducing Composite Evaluators in LangSmith
📊Combine multiple evaluator scores into one metric for a complete view of your app’s performance.
➕Supports weighted averages or weighted sums with configurable weights.
Learn more 👉 docs.langchain.com/langsmith/composite-evaluators -
LangChain
@LangChain
🧠🤖Deep Agents — now in JavaScript!
Simple tool-calling loops break down on long-horizon or intricate problems. Deep Agents, like Deep Research, Claude Code & Manus, chain reasoning, adapt plans, and juggle tools to get results.
Now you can build your own using our JS package! Links are below! -
LangChain
@LangChain
🔥 Our latest LangChain Academy course – Deep Research with LangGraph – is now live! 🔥
Deep research agents are taking off – from major AI labs to companies building their own.
Research is inherently open-ended. You can't always predict whether a question needs broad exploration or deep analysis. Agents excel here because they adapt on the fly, using each finding to decide where to dig next.
Building these systems ourselves and with customers, we've learned that structure matters. The best research agents scope problems with users first. Then, they coordinate multiple specialists instead of overwhelming one generalist.
LangGraph is built for these types of long-running, multi-agent workflows. Its persistence layer tracks progress across agents. LangSmith gives you the observability and evaluation tools you need to track and improve performance.
In this course, you'll build and evaluate:
- A user scoping agent to define research parameters
- A multi-agent research team with supervisor
- And add tool integration via MCP
Enroll for free ➡️ academy.langchain.com/courses/deep-research-with-langgraph/This media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
💻 Introducing Open SWE: An Open-Source Asynchronous Coding Agent
Open SWE is a fully autonomous, cloud-based coding agent that acts like a true teammate. Integrate it with your GitHub account, and put it to work fixing bugs or implementing features.
Try it out the demo today with just an Anthropic API key 👇 -
LangChain
@LangChain
🎙️🤖 podcastfy.ai
Transform text, images, websites & videos into multilingual audio conversations. Built with LangChain, it offers local LLM deployment and multi-speaker support for complete developer control.
Check out this open-source tool on GitHub 🚀
github.com/souzatharsis/podcastfy -
LangChain
@LangChain
Join @Redisinc and @langchain on July 23 for a webinar to learn how LangGraph + Redis make it easy to build AI agents with real memory. In the live webinar, you'll get demos, performance tips, and hear from the engineers behind the integration.
👉 RSVP here: -
LangChain
@LangChain
@ProsperLoans is a financial services company that connects everyday people with the financial solutions they need, from personal loans, credit cards, and more.
To automate tedious workflows like call QA, Zak Golding (GenAI Software Engineer) and the Prosper team built a modular AI agent platform using LangGraph. As a result, their team has been able to:
• Cut verification costs by over 90% — from dollars to cents per call
• Scaled from sampling calls to reviewing 100% of customer calls automatically
• Reuse agent components across teams and use cases -
LangChain
@LangChain
In healthcare, every decision must be grounded in accuracy and trust. @cityofhope — a national medical center specializing in cancer care and diabetes — developed HopeLLM to help with just that.
HopeLLM is a LangGraph-powered agent designed to help summarize patient journeys for physicians in real time. In just six weeks, the City of Hope team was able to:
• Save over 1,000 hours for clinicians by automating medical summaries
• Support 1,500+ unique patients across 2,300+ AI-assisted interactions
• Add in multi-layer validation and fact-checking to ensure clinical reliability -
LangChain
@LangChain
Our latest LangChain Academy course – Building Ambient Agents with LangGraph – is now available!
Most agents today handle one request at a time through chat interfaces. But as models have improved, agents can now run in the background – and take on long-running, complex tasks.
LangGraph is built for these “ambient agents,” with support for human-in-the-loop workflows and memory. LangGraph Platform provides the infrastructure to run these agents at scale, and LangSmith helps you observe, evaluate, and improve your agents. Together, they make it easier to build reliable, production-ready agents.
In this course, you’ll:
📩 Learn the fundamentals of LangGraph and how to build your own email agent
📊 Evaluate your agent using LangSmith
✅ Add human-in-the-loop for reviews, and memory so your agent adapts over time
🚀 Deploy your agent with LangGraph Platform and connect it to GmailThis media couldn't be displayed. See it on the original post. -
LangChain
@LangChain
📄🔍 GraphRAG Contract Analysis
A powerful solution combining GraphRAG and LangGraph agents to analyze legal contracts using Neo4j knowledge graphs, featuring benchmarks across multiple LLMs.
Check out the full implementation guide ➡️ towardsdatascience.com/agentic-graphrag-for-commercial-co… -
LangChain
@LangChain
Our open source Agent Chat UI now offers standardized support for file uploads!
Brace
@BraceSproul
Agent Chat UI now supports file uploads!
You can now upload images & PDFs in your chats, and have them automatically passed to your graph with no additional configuration! Try it out with your own graph today 🔗👇 -
LangChain
@LangChain
🦜🤖Introducing Open Agent Platform
An open-source, no-code agent building platform. OAP enables non-developers to build highly customizable agents, which connect to:
- 🛠️MCP Tools
- 📄LangConnect for RAG
- 🤖Other LangGraph Agents!
Try the public demo, or fork & customize it yourself! 🧵 -
LangChain
@LangChain
✨ Linkedin’s SQL Bot — powered by LangGraph and LangChain — makes data more accessible to everyone ✨
Data teams often spend too much time helping colleagues find the right data, creating bottlenecks that slow companies down.
💬 To solve this, LinkedIn built SQL Bot—an AI assistant that translates natural language questions into SQL, empowering teams to access the data they need independently. It finds the right tables, writes queries, fixes errors, and ensures employees can retrieve insights while maintaining the correct permissions.
🛠️ How it works: SQL Bot runs on a multi-agent system built with LangGraph and LangChain, providing the intelligence and flexibility needed for complex query generation.
LinkedIn joins industry leaders like Uber, Replit, Elastic, and Rippling in choosing LangGraph for its control and adaptability in building powerful AI agents. -
LangChain
@LangChain
We’re excited to share that LangChain has been named to the 2025 Forbes AI 50 list! Produced in partnership with Sequoia and Meritech Capital, the list spotlights the most promising privately-held AI companies in the world.
It’s an honor to be recognized alongside many of our customers—such as Clay, Mercor, and Writer—who are shaping the future of AI. -
LangChain
@LangChain
🤖🧠 Redis x LangGraph Memory
The new langgraph-checkpoint-redis integration enables AI agents with persistent memory, featuring thread-level preservation and cross-thread memory with vector search capabilities.
Enhance your AI agents with advanced memory - check it out! 👉 -
LangChain
@LangChain
Using `langgraph-checkpoint-redis`, you can bring @Redisinc's powerful memory capabilities to your LangGraph agents. Persist context across conversations with thread-level and cross-thread memory, plus built-in vector search.
Learn more: redis.io/blog/langgraph-redis-build-smarter-ai-agents-wit… -
LangChain
@LangChain
We’re excited to share that LangChain has been named to the 2025 Enterprise Tech 30 list!
It’s an honor to be recognized alongside many of our customers—Lovable, Unify, Clay, Mercor, Rippling, and Klarna—who are also shaping the future of enterprise tech.
Curated by over 100 leading VCs and corporate development leaders, the ET30 highlights the top private companies driving innovation in the enterprise space.
Huge thanks to @Wing_VC and @EricNewcomer—and congratulations to all the companies on this year’s list! -
LangChain
@LangChain
🧱Building AI apps using LangChain
A new series from our friends at @Redisinc on building AI applications using LangChain and Redis
Covers RAG and semantic caching
www.youtube.com/playlist?list=PL83Wfqi-zYZH8Boxt51keR4vZU… -
LangChain
@LangChain
🦜💬Introducing Agent Chat UI
An OSS web app for interacting with any LangGraph app (PY & JS/TS) via a chat interface
Agent Chat can render:
- Human/AI Messages
- Tool Calls/Results
- Human in the Loop
- Generative UI (tomorrow 👀)
Watch the video, and try it out today 👇 -
LangChain
@LangChain
🚀 LangGraph 0.3: Prebuilt Agents
We just released LangGraph 0.3, introducing prebuilt agents that streamline AI application development. These ready-to-use templates enhance LangGraph's core capabilities.
Learn more about prebuilt agents 📖
blog.langchain.dev/langgraph-0-3-release-prebuilt-agents/ -
LangChain
@LangChain
🚀 Announcing LangGraph v0.3 with Prebuilt Agents!
We're introducing LangGraph prebuilt and a collection of ready-to-use agent libraries built with LangGraph:
- LangGraph Prebuilt: high-level APIs for tool-calling agents (part of LangGraph)
- Trustcall for reliable structured extraction
- LangGraph Supervisor for supervisor multi-agent architectures
- LangMem for long-term memory
- LangGraph Swarm for swarm multi-agent architectures
We hope this will make it easy to get started with building agents and foster a large collection of prebuilt agents built by the community.
Want to contribute your own agents? Check out instructions in the blog (link below)! -
LangChain
@LangChain
🧠 LangMem: AI with Human-like Memory
We just launched an API SDK that enables AI agents to maintain memory across multiple conversations. Built with LangGraph, it makes AI interactions more natural and contextually intelligent.
👉 Discover the technical details: -
LangChain
@LangChain
🛠️ Introducing Langchain MCP Adapters: LangGraph agents 🤝 Anthropic Model Context Protocol
⚡ Instantly connect LangGraph agents to hundreds of tools in the growing MCP ecosystem!
Watch the video ➡️ youtu.be/OX89LkTvNKQ
Try it out in Python with `pip install langchain-mcp-adapters`
🐍 Repo: github.com/langchain-ai/langchain-mcp-adapters
🤖 MCP docs: modelcontextprotocol.io/introduction -
LangChain
@LangChain
🤖 🌐 WebRover 2.0: AI Web Co-Pilot
Introducing an autonomous web agent that revolutionizes online research with LangGraph. WebRover handles everything from task automation to academic research, complete with citations.
Explore this AI research companion! 🚀 -
LangChain
@LangChain
🗄️📊 Gen AI Toolbox for Databases
Google Cloud and LangChain just launched an open-source solution for secure database-to-AI connectivity. The toolbox streamlines AI agent development with LangGraph integration and robust authentication.
Start building secure AI agents today! 🚀
cloud.google.com/blog/products/ai-machine-learning/announ… -
LangChain
@LangChain
Worried about bias in your LLMs? See how @CVSHealth used LangChain to build LangFair, their Python library for assess fairness & bias and model responses.
The CVS Health eng team leveraged retrieval and memory to analyze biases over time.
Read more: medium.com/cvs-health-tech-blog/how-to-assess-your-llm-us… -
LangChain
@LangChain
💫 Why Uber, LinkedIn, Replit trust LangGraph for production-ready AI agents
Top companies rely on LangGraph to build reliable AI agents that power critical workflows in production. From Uber's AI-driven unit tests to LinkedIn's recruiter automation— LangGraph delivers control, flexibility, and efficiency at scale. -
LangChain
@LangChain
How Elastic evaluates AI agents with LangSmith
To improve their AI-powered Attack Discovery for threat detection, @Elastic implemented:
📊 Automated LLM evaluations: Shifting from manual checks to a robust, automated framework with LangSmith.
🔍 End-to-end observability: Tracing AI workflows using LangGraph + LangSmith for deep debugging.
🎯 Data-driven prompt tuning: Quantitative scoring to identify high-performing prompts and LLMs for specific tasks.
Read more: www.elastic.co/blog/elastic-security-generative-ai-features -
LangChain
@LangChain
⭐️Introducing the LangGraph Functional API
We spent a lot of time making LangGraph's state management (short/long term memory🧠, human in the loop🧑🤝🧑, time travel⏲️) production ready
You can now use these outside of the Graph syntax - with any framework (or none at all!)
With the release of the functional API, it is possible to get these features **without having to explicitly define a graph**
The functional API allows you to leverage LangGraph features using a more traditional programming paradigm, making it easier to build AI workflows that incorporate human-in-the-loop interactions, short-term and long-term memory, and streaming capabilities.
This API is complementary to the Graph API (StateGraph) and can be used in conjunction with it as both APIs use the same underlying runtime. This allows you to mix and match the two paradigms to create complex workflows that leverage the best of both worlds.
Blog: blog.langchain.dev/introducing-the-langgraph-functional-api/
Video: youtu.be/NXhyWJozM8A -
LangChain
@LangChain
🤖🌐WebRover AI Agent
Meet the open-source AI agent that autonomously navigates websites and automates tasks like a human
Powered by LangChain & LangGraph, it handles everything from research to finding the cheapest flights between cities.
github.com/hrithikkoduri/WebRover -
LangChain
@LangChain
At CES last night, Jensen announced NVIDIA blueprints to help accelerate agents in the enterprise
✨We're excited to collaborate with NVIDIA on a report generation blueprint using Llama 3 and LangGraph
This is a huge enterprise use case,
Check out the video walkthrough👇 -
LangChain
@LangChain
🌟 Case Study Spotlight: @DunBradstreet's ChatD&B is an AI assistant powered by LangChain & LangSmith, delivering trusted insights backed by data from 580M+ business entities.
ChatD&B supports workflows across credit risk, sales, supply chain, & more. Read their story to learn how they produced explainable, high-quality responses to help customers make data-driven decisions.
Read the Dun & Bradstreet story: blog.langchain.dev/customers-dun-bradstreet/ -
LangChain
@LangChain
🎉🎂 LangChain turns 2!
Two years ago, we launched a Python package called `langchain` to help developers build LLM applications that could reason. Today, we're still chasing that same mission — and as the generative AI ecosystem has evolved, we've evolved too.
LangChain has grown into a fuller ecosystem with LangSmith (for LLM testing & observability) and LangGraph (for building reliable agents). Our technology is now more stable, comprehensive, and production-ready.
✨ By the numbers, we've hit:
• 4,000+ contributors (doubled from last year)
• 130 million+ downloads and 132k apps powered by LangChain
• 700+ integrations (from just 3 providers at launch)
Thank you to our incredible partners, contributors, and LangChain community for being part of this journey. We're just getting started ❤️
➡️ Read our 2nd anniversary reflection: blog.langchain.dev/langchain-second-birthday/ -
LangChain
@LangChain
Learn how to create a RAG app from scratch using LangChain + @MongoDB's Python driver in MongoDB's latest course: "Using Atlas Vector Search for RAG Applications". See how to prepare data, integrate AI frameworks, and generate precise answers with Atlas Vector Search and LangChain.
Get started here 👉learn.mongodb.com/courses/using-atlas-vector-search-for-r…
See the course preview: learn.mongodb.com/learn/course/using-atlas-vector-search-… -
LangChain
@LangChain
🐘Long-term memory support in LangGraph🐘
LangGraph agents can now remember information across conversational threads, making it easier for them to adapt to users' needs.
Key features:
✅Cross-thread persistence of custom memories & content-based filtering through built-in document storage
✅Available for LangGraph Python and JS
✅Enabled by default for LangGraph Cloud & Studio users
Learn more:
📹 Conceptual memory overview: youtu.be/JTL0yp85FsE
📰 Blog post: blog.langchain.dev/launching-long-term-memory-support-in-…
Try it for yourself using our memory agent template:
📹 Tutorial: youtu.be/-xkduCeudgY
🐍 Python: github.com/langchain-ai/memory-agent
☕️ JavaScript: github.com/langchain-ai/memory-agent-js
We're excited to see what you'll build! -
LangChain
@LangChain
🤖 Data Enrichment Agent Template
Open-ended research -> Structured Output (CSV, DB) is tedious, making it well-suited for agents. Here, show a new LangGraph agent template for this, automating web research and writing to any user-supplied schema.
youtu.be/mNxAM1ETBvs -
LangChain
@LangChain
✨Announcing LangChain v0.3
Today, we’re excited to announce the release of LangChain v0.3 for both Python and JavaScript, focused on improved dependencies
Python: migrating to Pydantic 2
JS: moving to peer dependencies
Blog: blog.langchain.dev/announcing-langchain-v0-3/ -
LangChain
@LangChain
Honored to be recognized in the 2024 Intelligent Applications 40 as one of the top companies building and enabling AI applications. #ia40
Thank you @MadronaVentures, @AWSstartups, @Microsoft, @NYSE, @Delta, @McKinsey, & @MorganStanley. See the full list: www.ia40.com. -
LangChain
@LangChain
With the Box loader for LangChain, you can now easily integrate documents from @Box into your LangChain workflows.
You can store large amounts of unstructured data in Box and manage and process the data with LangChain. See how to implement the loader + reference potential use cases in the blog post below.
➡️Blog: medium.com/box-developer-blog/introducing-the-langchain-b… -
LangChain
@LangChain
🚨 JavaScript community, get ready for a week of exciting announcements! We’re kicking off with three major updates:
First, LangGraph.js 0.2 is live! LangGraph.js 0.2 is a controllable orchestration framework to build reliable agents in JavaScript/TypeScript. You can:
• Flexibly stream intermediate steps
• Rewind to and debug past states
• Add human-in-the-loop to approve actions
But that’s not all - LangGraph Studio (the first agent IDE) and LangGraph Cloud (our scalable deployment service for agents) also now support LangGraph.js applications.
Together, these form a powerful toolkit for JavaScript developers to prototype, debug, and deploy agents — with or without LangChain.
✍️ Read the blog post: blog.langchain.dev/javascript-langgraph-v02-cloud-studio/
📽️ Video introduction: youtu.be/Qu8BYTnh3K0
📓 LangGraph Studio starter repo: github.com/langchain-ai/langgraphjs-studio-starter
☁️ LangGraph Cloud docs: langchain-ai.github.io/langgraph/cloud/deployment/setup_j…
Stay tuned for more demos, tutorials, and videos all week! -
LangChain
@LangChain
We've teamed up with @cerebras to bring you lightning-fast LLM workflows ⚡️
With our latest integration, you can build and deploy faster than ever. Cerebras supports 1,800t/s for Llama3.1-8b and 450t/s for Llama3.1-70b.
Try out the Cerebras Module on LangChain: pypi.org/project/langchain-cerebras/
Check out the demo app: langgraph-cerebras.streamlit.app -
LangChain
@LangChain
🎉 We’re thrilled to announce our inclusion in Forbes' Next Billion Dollar Startups list! 🎉
The AI industry has made incredible progress over the past year with innovations in retrieval techniques, agent architectures, and new testing and evaluation methods. We’re grateful to be part of moving the industry forward and excited to see all of these startups recognized.
We’re actively hiring across various roles -- including Software Engineers, Deployed Engineers, and Account Executives -- come join our team!
➡ Careers: langchain.com/careers
➡ Forbes: www.forbes.com/sites/amyfeldman/2024/08/13/next-billion-d… -
LangChain
@LangChain
How Podium optimized agent behavior & reduced engineering intervention by 90% with LangSmith
Podium’s AI Employee - their flagship agentic application - helps small businesses boost lead conversion by providing fast, accurate responses to customer inquiries.
With LangSmith, @PodiumHQ could quickly create feedback loops to test their agent's performance - from dataset curation to model fine-tuning.
This resulted in:
• F1 scores surpassing Podium's 98% quality bar
• 90% reduction in engineering intervention due to more efficient issue resolution by the Support team
• Improved customer satisfaction (CSAT) scores for Podium's AI-powered services & support interactions
👉 Read Podium's story: blog.langchain.dev/customers-podium/
🤓 Check out our guide to testing LLM apps: www.langchain.com/testing-guide-ebook -
LangChain
@LangChain
And that's not all... LangGraph Cloud is now available in open beta. 🚀☁️
LangGraph Cloud lets you build fault-tolerant, scalable agents. You can:
• Manage large workloads with horizontally-scaling servers, task queues, and built-in persistence.
• Deploy agentic apps in one click, then debug and monitor associated traces.
• Use LangGraph Studio, a playground-like studio for debugging agent failure modes and quick iteration — now available also as a desktop app.
All LangSmith users on Plus and Enterprise plans can access LangGraph Cloud at no additional cost while in beta. Just navigate to the "Deployments" tab in your LangSmith UI to get started.
➡️ Sign up for a LangSmith account: smith.langchain.com
📓 Learn more in the docs: langchain-ai.github.io/langgraph/cloud/
🔁 ICYMI, LangGraph v0.2 is also released: blog.langchain.dev/langgraph-v0-2/ -
LangChain
@LangChain
🚀Announcing LangGraph Studio: The first agent IDE
LangGraph Studio offers a new way to develop LLM applications by providing a specialized agent IDE that enables visualization, interaction, and debugging of complex agentic applications
With visual graphs and the ability to edit state, you can better understand agent workflows and iterate faster. LangGraph Studio integrates with LangSmith so you can collaborate with teammates to debug failure modes
LangGraph Studio is available for free to all LangSmith users on any plan tier during its early development.
Read more about it here: blog.langchain.dev/langgraph-studio-the-first-agent-ide/
Watch a YouTube walkthrough: www.youtube.com/watch?v=pLPJoFvq4_M
Try out LangGraph Studio for free here: github.com/langchain-ai/langgraph-studio
Sign up for a LangSmith account: smith.langchain.com -
LangChain
@LangChain
🦙Llama 3.1 now in LangSmith Playground
Thanks to @FireworksAI_HQ for the fast integration! All Llama 3.1 models are now available in the LangSmith playground (for free!!!) for your immediate enjoyment :) -
LangChain
@LangChain
With the launch of the MongoDB AI Applications Program (MAAP) today, LangChain and @MongoDB are empowering enterprises to build, deploy, and scale AI-enriched applications.
We’re excited to be part of this ecosystem and to help enterprises build impactful LLM applications. #maaplaunch
💡 Learn more: www.mongodb.com/services/consulting/ai-applications-program -
LangChain
@LangChain
🤖everything-ai
Your fully proficient, AI-powered and local chatbot assistant
Let's you do RAG, summarization, audio translation, image generation... truly an all-encompassing local assistant
Fully OSS, and leverages lots of other OSS libraries!
github.com/AstraBert/everything-ai -
LangChain
@LangChain
🔒 We're excited to announce that LangSmith — our unified developer platform for building, testing and monitoring LLM applications — is now SOC 2 Type II compliant. After a rigorous audit process, LangSmith has been certified to conform to industry best-practices for the protection of data and for security procedures. 🔒
Security and compliance is important to us and to our growing list of enterprise customers. We're thrilled to be able to serve everyone more effectively.
➡ Sign up for LangSmith: smith.langchain.com
➡ Speak to a LangSmith expert: www.langchain.com/contact-sales
❓ Questions? Reach out to privacy@langchain.dev. -
LangChain
@LangChain
🎨🔧 Cohere tool calling 🎨🔧
You can now use tool calling with @cohere models in @langchain 🦜🔗 JS/TS agents + chains!
This improves reliability and adds support for standard methods like `withStructuredOutput`.
Try it in langchain/cohere 0.2.0:
js.langchain.com/v0.2/docs/integrations/chat/cohere/ -
LangChain
@LangChain
🚀 Introducing LangGraph Cloud 🚀
LangGraph helps you build reliable agents that actually work. Today, we've launched LangGraph Cloud, our new infrastructure to run fault-tolerant LangGraph agents at scale.
With LangGraph Cloud, you can:
• Handle large workloads with horizontally-scaling servers, task queues, and built-in persistence
• Debug agent failure modes and iterate quickly in a visual playground-like studio
• Deploy in one-click and get integrated tracing & monitoring in LangSmith
This builds on LangGraph v0.1, our latest stable release supporting diverse control flows - whether it's single- or multi-agent, hierarchal or sequential. You'll be able to create flexible agents with human-in-the-loop collaboration and first-class streaming support.
We can't wait to see the agentic workflows you'll ship. LangGraph Cloud is available in closed beta today.
👉 Join the waitlist for LangGraph Cloud: www.langchain.com/langsmith/deployment
✍️ Read our blog post announcement: www.langchain.com/blog/langgraph-cloud
📽️ Watch the video walkthrough: www.youtube.com/watch?v=l4sMKF1dTDM
🌀 Haven't tried LangGraph yet? Give it a spin here: langchain-ai.github.io/langgraph/
🙌 Check out the new LangGraph landing page: www.langchain.com/langgraph -
LangChain
@LangChain
LangChain has been named to @Redpoint’s 2024 InfraRed 100, celebrating the most transformative infrastructure companies! Check out the whole list and the InfraRed report here: www.redpoint.com/infrared/report/
-
LangChain
@LangChain
LangChain is the 2024 Databricks GenAI Partner of the Year! We are honored and grateful to collaborate with @Databricks in advancing outcomes for our shared customers.
See the blog post:
databricks.com/blog/databricks-announces-2024-global-part… -
LangChain
@LangChain
📷 Generative UI with LangChain Video Series
We're releasing a three-part video series on building generative UI applications with LangChain. The intro video is out today — see the inner workings of how the gen UI API works, use cases, and ways it can unlock new user experiences.
Watch the video here: youtu.be/mL_KuQgX9Oc
Stay tuned for more videos coming to the LangChain YouTube channel, where we'll dive deeper into building apps with GenUI: www.youtube.com/@LangChain -
LangChain
@LangChain
Recipes for open source / local agents w/ Llama 3
Since the release of Llama 3, we've seen high interest in building agents that run reliably and / or locally (e.g., on your laptop). We've partnered with @meta to release several new recipes for Llama 3 agents using LangGraph.
1/ We show how to build LangGraph tool-calling agents, using @groq, which supports fast inference and tool-calling.
2/ We show how to RAG agents using LangGraph and @groq, capable of complex self-corrective control flow.
3/ We show how to take the RAG agent and make it run locally using @nomic_ai embeddsm, and @ollama.
➡️ Code:
github.com/meta-llama/llama-recipes/tree/main/recipes/use…
📽️ Video:
www.youtube.com/watch?v=j2OAeeujQ9M -
LangChain
@LangChain
Excited to be in good company 🤝
Pinecone
@pinecone
The Pinecone Partner Program is expanding, and we are honored to welcome: @apivy, @1contextdata, @EstuaryDev, @flowiseai, @Gathr, @honeyhiveai, @instill_tech, @langchain, @octoaicloud, @pulumicorp, @streamnativeio, & @unstructuredio!
-
LangChain
@LangChain
🛝Enter the playground in LangSmith from scratch instead of entering from a run or a prompt 🛝
Playground is now available as its own tab in the sidebar of LangSmith. To create a new prompt, simply craft a prompt in the empty playground and click "Save As" to name it. This streamlines prompt creation and allows for playground experimentation with a straightforward entry point.
📓 Docs: docs.smith.langchain.com/how_to_guides/prompts/create_a_p…
🌟 Try it out in LangSmith: smith.langchain.com/playground -
LangChain
@LangChain
💻GPT Computer Assistant
An open source effort to replicate the ChatGPT desktop app!
You can easily install via pip, and then send text & screenshots to it and get back a voice response
The speed of the OSS community is incredible
github.com/onuratakan/gpt-computer-assistant -
LangChain
@LangChain
🧠Cognita
Cognita is an open source framework for building modular RAG applications
It builds on top of LangChain's low-level abstractions to provide a more opinionated and out-of-the box experience for RAG
We love to see frameworks like this!
github.com/truefoundry/cognita -
LangChain
@LangChain
🔒 LangSmith is now GDPR compliant! 🎉
We're excited to announce that LangSmith — our unified developer platform for building, testing, and monitoring LLM applications — is now GDPR compliant. Security and compliance is important to us and to our growing list of enterprise customers. We're thrilled to be able to serve everyone more effectively.
➡ Sign up for LangSmith: smith.langchain.com
➡ Interested in our Enterprise plan? Speak to a LangSmith expert: www.langchain.com/contact-sales -
LangChain
@LangChain
🤝Hugging Face x LangChain partner package
We're excited to announce the launch of langchain-huggingface, a partner package in LangChain jointly maintained with @huggingface.
LangChain users can now reliably connect to and access Hugging Face features. These include chat, text completion, and embedding models for both local & hosted instances.
Read the blog post: huggingface.co/blog/langchain
... And stay tuned for more to come 😉 -
LangChain
@LangChain
🚀 Pre-release of LangChain v0.2
The pre-release of LangChain v0.2 is out today, paving the way towards greater stability + security.
Read the full blog here: blog.langchain.dev/langchain-v02-leap-to-stability/
With the help of our users and developers, we've iterated on our stable library version and have decoupled langchain from langchain-community to make langchain more lightweight, focused, and secure.
With v0.2, you can look forward to:
• Versioned docs with improved discoverability, so you can match information and instructions with your minor version
• LangGraph as the recommended way to build agents, which makes it easy to define cycles and built-in memory — and then customize and modify them
Since our last v0.1 release, we've also made improvements to:
• Standard chat model features, including standardized tool calling support and an interface for structuring output
• Async support for many core abstractions
• Streaming support with a new Event Streaming API
• Partner packages for 20+ providers across Python and JavaScript packages
Thank you to our community for helping us build better and more reliably! 🤗
👂Give your feedback: github.com/langchain-ai/langchain/discussions/21437 -
LangChain
@LangChain
🎉 We’re thrilled to be MongoDB's AI App Framework Partner of the Year!
LangChain is the #1 choice for developers building GenAI apps. We have immense gratitude for our community of over 2,000 contributors who have helped us shape the LLM application development space, and are humbled to support over 50,000 companies who choose to build with LangChain!
www.mongodb.com/blog/post/building-the-next-big-thing-tog… -
LangChain
@LangChain
Adaptive RAG w/ Cohere's new Command-R+
Adaptive-RAG (@SoyeongJeong97 et al) is a recent paper that combines (1) query analysis and (2) iterative answer construction to seamlessly handle queries of differing complexity.
We took at stab at implementing these ideas from scratch using a ReAct agent and LangGraph with @cohere's Command-R and the new Command R+.
Command-R is fast and lightweight (35b parameter) with strong tool-use and RAG performance. It works very nicely w/ LangGraph, performing query analysis (re-writing and routing) between a vectorstore, web search, and fallback to LLM. We also perform RAG with fast in-the-loop unit tests for doc relevance, answer hallucinations, and answer quality.
We show the same same workflow using a ReAct agent and the larger Command R+. In the video, we discuss the trade-offs between using agents vs LangGraph, and Command-R vs the newer / larger Command R+.
Video:
youtu.be/04ighIjMcAI
LangGraph code:
github.com/langchain-ai/langgraph/blob/main/examples/rag/…
ReACT agent code:
github.com/cohere-ai/notebooks/blob/main/notebooks/react_…
Paper:
arxiv.org/abs/2403.14403 -
LangChain
@LangChain
👂 @cohere Integration Package 🦜
Hot off the heels of Cohere's v5 sdk supporting tool calling and their new Command R+ model, we're excited to announce the `langchain-cohere` package, which exposes all of their integrations ranging from their chat models to model-specific agents.
Py Docs: python.langchain.com/docs/integrations/providers/cohere
JS Docs: js.langchain.com/docs/integrations/chat/cohere -
LangChain
@LangChain
@AI21Labs Foundation and Task-Specific Models 🎯
Unlock the full potential of AI21's Jurassic-2 and Task-Specific models directly in LangChain! Effortlessly build and scale generative AI applications that meet today's business needs. No NLP expertise required; just seamless integration for text generation and comprehension. Get started now and transform your projects with the power of AI21 and LangChain.
python.langchain.com/docs/integrations/chat/ai21 -
LangChain
@LangChain
✨ Today, we’re thrilled to announce ✨
- The general availability of LangSmith (no more waitlist!)
- Our Series A fundraise led by @sequoia
- Our beautiful new homepage and brand
We've worked hard over the past few months to add requested features and ensure LangSmith can operate at scale. We’re now confident in saying that it is the most complete platform for building production-grade LLM applications, whether or not you’re using LangChain. Learn more here: blog.langchain.dev/langsmith-ga/ and sign up here: smith.langchain.com
Our series A round will give us the capital needed to grow our open source and platform offerings. Working with @sonyatweetybird, @romie_boyd, and the rest of the Sequoia team has been a privilege so far! www.forbes.com/sites/alexkonrad/2024/02/15/open-source-ai…
Finally, we’re excited to unveil our new homepage and brand. Dive into our new website at www.langchain.com to see the changes for yourself, explore the expanded resources, and discover what LangChain, LangSmith, and LangServe have to offer.
PS — we’re hiring! Explore our careers page and reach out if you think you’re a fit for any of our open positions! www.langchain.com/careers -
LangChain
@LangChain
🤖BCG X Releases AgentKit, a Full-Stack Starter Kit for Building Constrained Agents
AgentKit is a LangChain-based starter kit to build constrained agents, developed by our partners at BCG X
We've often found that in order to productionalize agentic applications, you need to develop them in a fairly constrained way. This will help with that!
This is a full stack application, built on NextJS, FastAPI, and LangChain
BCG X has already used this to develop:
🧪Generating drafts of complex clinical documents, such as clinical trial protocols, for a global pharma company
🚢Controlling and orchestrating supply chain optimization systems using a helpful agent assistant
🚗Developing a chatbot that helps a major automotive player service its customers
Check out the code here: github.com/BCG-X-Official/agentkit
Read the full blog here: blog.langchain.dev/bcg-x-releases-agentkit-a-full-stack-s… -
LangChain
@LangChain
🌲RAG With PDFs
Want to easily set up a RAG pipeline over PDFs?
This great tutorial by our friends at @clarifai goes over how to do exactly that
Blog: www.clarifai.com/blog/ai-in-5-rag-with-pdfs -
LangChain
@LangChain
📹Just finished up our webinar with @plasticlabs which dove deep on memory.
some interesting similarities between their paper and the Generative Agents paper:
-both used a vectorstore as their datastructure
-but both stored reflections (using an LLM to generate insights from raw events) rather than just raw events
even though using a vectorstore may seem simple, its more about what goes in the vectorstore
www.youtube.com/watch?v=PbuzqCdY0hg -
LangChain
@LangChain
📅 7/17 will be the biggest LangChain meetup yet
🦜Connecting LLMs to your data
Featuring:
@drewhouston @Dropbox
@hwchase17 @langchain
@Bahaaldine @elastic
@chipro @ClaypotAI
@karinanguyen @AnthropicAI
3⃣cities (live in SF, streaming in Seattle and NYC)
signup 👇 -
LangChain
@LangChain
LangChain ❤️ @OctoML
OctoAI is world-class compute infrastructure for tuning and running models that wow your users.
We recently added integrations for LLM and embedding models on their platform!!
integrations.langchain.com/llms











































































































































































































































































































































































































