Repolode
Submit
langchain-ai

langchain-ai/langgraphjs

A TypeScript framework for building stateful, long-running LLM agents as graphs, with persistence, memory and human-in-the-loop.

View on GitHubWebsite
Stars
3,371
Forks
610
Open issues
174
Last commit
7 hours ago

What is it?

LangGraph.js is the JavaScript and TypeScript version of LangGraph, a low-level orchestration framework for building agents. Developers define an agent as a graph of nodes and edges that operate on shared state, which gives them explicit control over the flow of reasoning, tool calls and branching instead of relying on a fixed chain.

The framework targets workflows that must survive failures and run for extended periods. Execution is durable, so an agent can resume from where it stopped. It supports short-term working memory within a session and long-term memory across sessions, and it lets a human inspect or modify agent state at any point during a run.

LangGraph.js is published on npm as @langchain/langgraph and works alongside @langchain/core. It can be used on its own, and it is designed to pair with LangChain for integrations and with LangSmith for tracing and evaluation.

Key features

  • 01Defines agents as graphs of nodes and edges over shared state.
  • 02Durable execution that resumes automatically after failures.
  • 03Short-term and long-term memory for stateful agents.
  • 04Human-in-the-loop inspection and modification of agent state during runs.
  • 05Written in TypeScript and installed from npm.

Use cases

  • Building multi-step agents that need explicit control over branching and tool calls.
  • Running long-lived agent workflows that must resume after crashes or restarts.
  • Adding human approval steps to agent actions before they execute.
  • Building JavaScript or TypeScript backends that orchestrate LLM calls.

Quick start

  1. 1

    Install the LangGraph and LangChain core packages with npm

    npm install @langchain/langgraph @langchain/core

Things to know

  • Low-level framework, so building agents requires more code than higher-level tools such as Deep Agents.
  • Some features like the best debugging and deployment experience rely on the separate LangSmith product.

Topics

#agents#llm#typescript#orchestration#workflow#memory#human-in-the-loop#ai#artificial-intelligence#generative-ai

FAQ

What is langgraphjs?
A TypeScript framework for building stateful, long-running LLM agents as graphs, with persistence, memory and human-in-the-loop. LangGraph.js is the JavaScript and TypeScript version of LangGraph, a low-level orchestration framework for building agents. Developers define an agent as a graph of nodes and edges that operate on shared state, which gives them explicit control over the flow of reasoning, tool calls and branching instead of relying on a fixed chain.
Is langgraphjs free to use? What license does it have?
langgraphjs is open source under the MIT license. Check the license terms before using it in commercial products.
Is langgraphjs still maintained?
Actively maintained. The last commit to langgraphjs was on Oct 21, 2026.
What can I use langgraphjs for?
Building multi-step agents that need explicit control over branching and tool calls. Running long-lived agent workflows that must resume after crashes or restarts. Adding human approval steps to agent actions before they execute. Building JavaScript or TypeScript backends that orchestrate LLM calls.
How do I install langgraphjs?
Install the LangGraph and LangChain core packages with npm: npm install @langchain/langgraph @langchain/core
Where can I get langgraphjs?
The source code and installation instructions are on GitHub: https://github.com/langchain-ai/langgraphjs

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects