📚🔍 Event Deep Research An AI system that transforms historical research into structured timelines, automatically extracting and organizing biographical data from multiple sources into chronological JSON format. Check out the project 🎯 https://lnkd.in/gUY7PbJb
About us
LangChain is the platform for building reliable agents. Our products power top engineering teams — from fast-growing startups like Lovable, Mercor, and Clay to global brands including AT&T, Home Depot, and Klarna. LangGraph is a low-level orchestration framework for building controllable agents and long-running workflows. It’s used in production by teams at Replit, Uber, LinkedIn, GitLab, and more. LangSmith offers unified evaluation and monitoring to help developers debug, evaluate, and improve their agents at scale. LangChain provides hundreds of integrations and composable components, making it easy to connect with the latest models, tools, and databases — with minimal engineering overhead. Together, these tools help teams build, deploy, and manage enterprise-grade agents, faster.
- Website
-
langchain.com
External link for LangChain
- Industry
- Technology, Information and Internet
- Company size
- 51-200 employees
- Type
- Privately Held
Products
![Click here to view LangChain]()
LangChain
Software Development Kits (SDK)
LangChain is the platform for building reliable agents. Our products power top engineering teams — from fast-growing startups like Loveable, Mercor, and Clay to global brands including AT&T, Home Depot, and Klarna.
Employees at LangChain
Updates
-
🧠 LangGraph × cognee Integration cognee brings persistent memory to LangGraph agents, letting AI applications maintain context across sessions while seamlessly working with existing LangGraph features. Check out how to add memory to your agents 🔗 https://lnkd.in/gAUc2S_s
-
-
Launch Week is right around the corner — and we're kicking it off in person! We'll be hosting meetups in San Francisco, Boston, and NYC to celebrate LangChain's 3rd birthday and share what's coming next during Launch Week. Come hang with the team, connect with our amazing community of builders shipping agents in production, and get the inside scoop on what's happening during Launch Week. Expect great food, drinks, and good conversation. It’s going to be an exciting evening! 📍 Pick your city and RSVP: 🌉 SF: https://luma.com/7baj9rx5 🦪 Boston: https://luma.com/135zbg4u 🗽 NYC: https://luma.com/f5jrv7t6 See you there 👋
-
In this blog piece, you’ll learn why and how we built LangGraph for production agents. Building upon feedback from the super popular LangChain framework, we aimed to find the right abstraction for AI agents, and decided that was little to no abstraction at all. Instead, we focused on control and durability, and the core features needed to scale. https://lnkd.in/gaY9gHVH
-
LangChain reposted this
FOUNDING ENGINEER: probably the most important and under-discussed role at an early stage startup. I sat down with founding engineers from LangChain (Nuno Campos), Perplexity (Nikhil Thota), and Stytch (Alex Zaldastani) to talk about what makes a good one, what they look for, how they make technical decisions, and more:
-
Learn how to use LangSmith to debug your AI applications. 🛠️ In this video, we’ll show you how "Studio," our IDE for building agents, works, and how you can use it with any LangGraph agent you’ve built. You’ll also get everything you need to get started with Studio. https://lnkd.in/duhc72Qr
Getting Started with LangSmith (3/8): Debugging with Studio
https://www.youtube.com/
-
LangChain reposted this
As your organization builds agentic AI, one architectural imperative stands out: 𝗵𝘂𝗺𝗮𝗻-𝗶𝗻-𝗹𝗼𝗼𝗽 𝗰𝗼𝗻𝘀𝗲𝗻𝘁 + 𝗮𝘂𝘁𝗵𝗼𝗿𝗶𝘇𝗮𝘁𝗶𝗼𝗻. Agents need dynamic access, not permanent keys. Langgraph enables exactly this human-in-loop consent flow: agents interrupt, ask for permission, then resume. 𝗪𝗵𝘆 𝗵𝘂𝗺𝗮𝗻-𝗶𝗻-𝗹𝗼𝗼𝗽 𝗶𝘀 𝗻𝗼𝗻-𝗻𝗲𝗴𝗼𝘁𝗶𝗮𝗯𝗹𝗲 𝗳𝗼𝗿 𝘁𝗵𝗲 𝗲𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲: 🌱 Prevent privilege creep - agents only gain new permissions when a human explicitly approves 📜 Ensure auditability - every access request and consent step is logged, reviewed, and reportable 🛡️ Build governance confidence - executives, compliance, and security teams can see oversight in action 🔒 Least privilege + just-in-time - agents receive only the access they need, when they need it 𝗛𝗼𝘄 𝗟𝗮𝗻𝗴𝗚𝗿𝗮𝗽𝗵 / 𝗟𝗮𝗻𝗴𝘀𝗺𝗶𝘁𝗵 𝗱𝗲𝗹𝗶𝘃𝗲𝗿: 1. You register OAuth providers (GitHub, internal APIs, etc.). 2. When the agent needs new access, it pauses and surfaces a consent URL for the human to approve. 3. On consent, the agent resumes with the scoped token - future flows are smoother thanks to token refresh logic. 4. Optionally, tokens can be human-scoped (shared across agents) rather than tied to a single agent. Check out link to blog on Agent Authorization and docs link on how to set up Agent Auth in Langgraph.
-
-
We're hosting an intimate evening in Boston to celebrate LangChain's 3rd birthday and share our biggest product releases of the year. If you're an AI builder and want to connect with the community (plus get an early look at what we've been working on), join us. 👉 RSVP: https://luma.com/135zbg4u
-
-
Learn about the different types of runs you can create while tracing in LangSmith—and how they help you understand your application’s execution. Unlike traditional logs (or traces), which can be hard to interpret for LLM applications, LangSmith makes it easy to explore and debug with an intuitive, purpose-built UX. https://lnkd.in/e5u-Sgar
Getting Started with LangSmith (2/8): Types of Runs
https://www.youtube.com/
-
We just added a long awaited guardrails guide to the LangChain docs! Ship safer agents faster with: 🛡️ Built-in PII redaction 🫅 Human-in-the-loop approvals Or build custom guardrails that trigger before/after your model calls. https://lnkd.in/eWWFH8PB