Next stop on the LangSmith Roadshow? Chicago! Hear about the current state of agent development, and get ready for dive into two hands-on workshops: Agent harnesses with Deep Agents Explore the core concepts behind agent harnesses and learn how to build with Deep Agents. Improving your agents with LangSmith Engine Get hands-on with an agent for agent engineering that identifies where your agent is failing and helps make the fix for you. This event is designed for agent engineers, AI platform teams, and technical leaders building and operating agents in production. Interested in attending? RSVP below ⬇️ https://lnkd.in/eDxQPaNZ
LangChain
Technology, Information and Internet
San Francisco, CA 559,122 followers
Own your intelligence.
About us
LangChain enables every company to own their intelligence. LangChain's open, model-agnostic harnesses give teams choice and control over how they build their agent architecture. LangSmith brings testing, deployment, and monitoring together so teams can continuously improve their agents, compound intelligence, and govern them at scale. More than 7,000 customers, including Nvidia, Bridgewater, LinkedIn, Workday, Harvey, and Rippling trust LangSmith to improve their agents across the Agent Development Lifecycle.
- Website
-
langchain.com
External link for LangChain
- Industry
- Technology, Information and Internet
- Company size
- 201-500 employees
- Headquarters
- San Francisco, CA
- Type
- Privately Held
Employees at LangChain
View 469 employees at LangChain
or
Locations
Updates
-
We have several LangChain-led and community-led meetups coming up over the next few months! 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
-
-
-
-
-
+6
-
-
Join us, Extend, and Ramp this Thursday for a private screening of You Can See Everything in San Francisco. Seats are limited, save yours now! https://luma.com/27b59c8i
-
-
Join us in Seattle for our first ever Seattle Meetup!
It's our first ever Seattle LangChain Meetup, and we're going big! Join us on November 18th, reserve your spot now - https://luma.com/h3w0aaol The event will be highlighted by a presentation from LangChain Co-Founder/CEO Harrison Chase, and a sit-down conversation between Harrison and Expedia Group's Chief AI & Data Officer Xavier (Xavi) Amatriain . Special thanks to Mani Najmabadi for hosting us at Expedia Group's Seattle campus. This is a special opportunity to hear from visionaries leading the agentic world. Excited to have the LangChain developer community show out. See you there!
-
On October 28th, you're invited to 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: https://lnkd.in/e_6EJSTh
Our design agent at LangChain has shipped over 3,000 assets and posts with its work pulled more than 4.5M views on socials. It made the promo for our next webinar too 🦜 It’s become the design brain for the whole company. Decks, diagrams, ad creatives, launch films and more, all requested in Slack. In fact, three in four requests now come from outside marketing! Some context on why we built it. LangChain runs like a media company now, with constant product launches, campaigns, webinars and events, and every one of them needs design work. Our design team was stretched thin with repetitive requests, and as a lean startup we needed a way for anyone at the company to make high-quality, on-brand work we'd be proud to ship. We encoded our brand designers’ judgment into the agent’s skills, a design wiki and brand checks that run on every render. Then we built a custom media harness around it: a headless Photoshop, a shared visual canvas with Paper, video tools like Hyperframes from HeyGen plus our own animation engine, our brand assets and a library of design references. It all runs on Managed Deep Agents, which abstracts away sandboxes, memory, the Slack integration and schedules. Our own code shrank by about 80% after we migrated to it, so we spend our time on the design work instead of the infrastructure! On Oct 28 at 11am PT, Shreya Talegaonkar , Jan Franco Glanc Gomez, Daniel Lambert I are walking through the architecture, use cases and impact at our GTM Engineering Live! The full blog post and the open-source repo drop next week too. Register here: https://lnkd.in/g7P2Kwsy
-
NOW ON YOUTUBE (link below): Guess I am a speaker now? (Not sure about that.) But I learned so much from speaking at LangChain Interrupt 2026 last week in London - but not what you think?! Preparation of this talk reminded me a lot of my days (or years) running experiments in bio labs 🧪 👨🔬 more than a decade ago. (doing cancer research btw) I again started to think of our work as a journey leading to a story truely worth telling. Designed experiments precisely around gaps in the talk's argumentation flow... Every enterprise needs an AI lab RIGHT NOW - people finding those stories worth telling. This can be a one-person job, or a small team's. And it needs to be independent of the production pipeline. Uncovering immensely valuable use cases nobody dares to even think about. Model capabilities already exceed what many of us deem possible. SOTA models like Astra or Opus 5.5 show visual 2D/3D understanding that is getting much closer to how we humans see the world. So we at DB Engineering & Consulting dared to put Astra to the test. Watch my talk on how a vision deep agent converts hand-drawn railway plans to perfect CAD files using only 2 core tools 🧰: https://lnkd.in/dSYK9VGn
-
-
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.
Today we’re introducing Argos, Corridor’s first security model, to steer coding agents toward writing more secure code. Argos reads a coding agent’s plan and selects codebase-specific security guidance before feature implementation. These requirements become part of the plan the agent uses to write the code, which means more secure code by default, and less vulnerable code for developers and security teams to fix later. We fine-tuned a 9-billion-parameter model for this task using Baseten Loops. Argos exceeds the frontier at cost and speed. It had the highest combined precision and recall score of the 14 models we tested. Argos is lightning-fast: it returns a complete selection in 2.5 seconds on average, to avoid developer friction. The post covers how we built Argos, evaluated its selections, and used LangSmith to inspect the security context agents receive in production. Learn more here: https://lnkd.in/gpRuqfzh
-
-
We're partnering with Quantiphi to help even more organizations own their intelligence. Full announcement here ⤵️
Quantiphi and LangChain have announced the partnership and are set to transform how enterprise AI works. Historically, enterprise AI has been a consumption model where you pay for outputs but retain no ownership over the intelligence. We are shifting that paradigm. Quantiphi helps organizations own their intelligence and compound their competitive edge through sovereign AI foundations. You keep absolute control over your data, IP, and infrastructure portability. This partnership pairs LangChain’s powerful technical foundation with Quantiphi's forward-deployed engineering to close the implementation gap, and that combination is what gets enterprises from pilot to production, for real. Explore our joint capabilities: https://lnkd.in/d_w6sPga #Quantiphi #SolvingWhatMatters #LangChain #LangSmith #Partnership #SovereignAI #AgenticAI #AIEngineering #EnterpriseAI Asif Hasan | Sidd Kotwal | Harrison Chase | Darren Moffett | Karan Singh | Nithin Bose | Ravi Teja Konkimalla
-
Agents cram for exams too. The more history and tool output a model carries, the worse its reasoning gets. Deep Agents handles this in two ways: 1️⃣ Offloading. When a tool result comes back huge, like a 20,000 token database query, it’s saved to a virtual file system and the model only sees a preview. It can page through the full result with the read file tool if it needs to. 2️⃣ Summarization. Once the context window hits 85% (a default you can change), older messages get condensed into a summary so the agent has room to keep working. Nothing is deleted. A quick explanation from Jess Weng Want to learn more? Sign up for Agent Memory 101: Context Engineering https://lnkd.in/eutUvUUh
-
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. 🌏 Register: https://lnkd.in/gyKpP7Ez
