We’re thrilled to announce cross-thread memory support in LangGraph, available for both Python and JavaScript. This new feature lets your agents store and recall information between conversation threads for smarter, more adaptive interactions.

Key Benefits of Cross-Thread Memory:

• Persistent Memory: Store and recall information across different conversation sessions.

• Flexible Namespacing: Organize memories using custom namespaces for different users, organizations, or contexts.

• JSON Document Storage: Save memories as JSON documents, making it easy to manipulate and retrieve data.

• Content-Based Search: Easily filter and search memories across namespaces based on content.

Why Cross-Thread Memory?

Most AI agents today lose context between conversations, severely limiting their potential. LangGraph now changes this by providing a simple, reliable, persistent memory layer that works across threads, allowing your agents to learn from user feedback, retain preferences, and continuously improve.

Getting Started

• Conceptual guides: Explore the concept of memory in LangGraph for both Python and JavaScript.

• How-to guide: Get step-by-step instructions for sharing memory across threads in your LangGraph applications. See the Python guide and JavaScript guide

• Implementation Tutorials:

  • An end-to-end tutorial video walking through the implementation

  • A LangGraph Memory Agent showcasing a LangGraph agent that manages its own memory

  • A LangGraph.js Memory Agent to go with the Python version

• To run memory tasks in the background, we've also added a template and video tutorial on how to schedule memory updates flexible and ensure only one memory run is active at a time.

Explore these materials to leverage long-term memory in your LangGraph projects today!