WFH retweeted
☕ Now available in TypeScript!
LangGraph.js docs now include guides for building agents — covering memory, human-in-the-loop, streaming, deployment, and more!
Check out the link below 👇
WFH retweeted
It can be helpful to pinpoint the in-context information that a language model uses when generating content (is it using provided documents? or its own intermediate thoughts?). We present Attribution with Attention (AT2), a method for doing so efficiently and reliably! (1/8)
WFH retweeted
You can find the full blog here: blog.langchain.dev/how-to-th…
You can find a spreadsheet comparing agent frameworks here: docs.google.com/spreadsheets…
I tried to be as specific and nuanced as possible. I would love any feedback!
WFH retweeted
🚀 New docs drop!
LangGraph docs now include guides for building agents — covering memory, human-in-the-loop, streaming, deployment, and more.
🧵 A quick tour of what's covered:
WFH retweeted
🧹By popular demand, today we've added `delete_thread(thread_id)` method to all LangGraph checkpointers. This method deletes all data associated with the thread_id passed in.
langchain-ai.github.io/langg…
WFH retweeted
🔥🦜 Custom HTTP routes and middleware for LangGraph.js
Extend your LangGraph.js deployment with a @honojs app and build anything from a simple webhook handler to a full-blown backend for your API app.
WFH retweeted
I'm so excited to be speaking @Microsoft Build for the first time this year🥳 I'll be giving a workshop on building a Deep Researcher with Langchain (LangGraph to be specific!)
See you in Seattle 🎉🎉
build.microsoft.com/en-US/se…
WFH retweeted
🕸️Better stack traces in latest LangGraph Python!
We've done some work to remove internal frames from the stack traces when there is an error in your graph's nodes or edges. This makes it easier to find the error in your code without distractions
WFH retweeted
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.
WFH retweeted
🔒🦜Custom Authentication and Authorization for LangGraph.js
Add production-grade auth into your LangGraph deployments, no other backend or proxy required.
Now available for LangGraph.js in Cloud and self-hosted environments!
WFH retweeted
Custom auth for LangGraph.js deployments!!!
I know people toss around the term "game changing" a lot, but this seriously is a game changing improvement for LangGraph.js!!
Huge shoutout to David
WFH retweeted
I really feel like we're trending for the most boring timeline, where Google just wins everything (as anyone would've guessed 10 years ago):
- incredible new ml hardware
- Gemini 2.5 crushing it left and right
- waymo gonna sweep self driving taxi market
...
WFH retweeted
Massive impact. Open source. Let’s go! (Remote or in person!) 🚀💥🧩
This quoted post is unavailable.
WFH retweeted
🚀Over 2x speed improvement for checkpointing in LangGraph Python 0.3.27 out now
Introducing new invoke/stream arg, checkpoint_during=False, to checkpoint only the last step in each run.
Read on 🧵
WFH retweeted
🚀LangChain Perplexity Integration
Perplexity access is out! Easily install with `pip install langchain-perplexity` to get started with research oriented tasks in your applications.
This is an important step as we move to add support for citations and deep research tools 🤓.
WFH retweeted
🚀 New batch of performance improvements for LangGraph Python, out now
Up to 1.4x faster on our benchmark graphs
Read on 🧵
WFH retweeted
🎨 Native Generative UI for LangGraph Python
Colocate and serve your React UI components with your LangGraph agent. Now supported in Python alongside JS!
WFH retweeted
🧹TTLs in LangGraph Platform!
You can now configure your LGP deployments to clean up old threads or memory items automatically in the background
Read on 🧵
WFH retweeted
⚙️🤝 Customizable handoffs in LangGraph Supervisor!
You can now create custom handoff tools, giving you precise control over how the supervisor assigns tasks to individual agents. Customize tool names, descriptions, and data passed to worker agents!
Try it out with `pip install -U langgraph-supervisor`
























