langgraphjs

Active
GitHub TypeScript MIT

Description

Framework to build resilient language agents as graphs.

Key Features

  • Durable execution - Agents persist through failures and resume from exactly where they left off
  • Human-in-the-loop - Seamlessly incorporate human oversight by inspecting and modifying agent state at any point
  • Comprehensive memory - Both short-term working memory for ongoing reasoning and long-term persistent memory across sessions
  • LangSmith debugging - Deep visibility with execution path tracing, state transition capture, and runtime metrics
  • Production-ready deployment - Scalable infrastructure designed for stateful, long-running workflows
  • Graph-based orchestration - Low-level directed graph agent orchestration with customizable architectures

Use Cases

💡 Build complex AI agent workflows requiring long-running execution and fault recovery
💡 Introduce human oversight at critical decision points for reliable and safe agent behavior
💡 Define multi-step agent logic using graph structures with loops, conditional branching, and parallel execution
💡 Integrate with LangChain ecosystem and use LangSmith for deep debugging and observability

Categories

Quick Start

Install npm install @langchain/langgraph @langchain/core, define your agent graph with createGraph(), compile and invoke. Use @observe() for tracing. Check out LangChain Academy free courses for structured learning and the streaming cookbook for advanced patterns.

Related Projects