# LangGraph

LangGraph is an orchestration framework and runtime for building long-running, stateful agents and controlled agent workflows.

Canonical URL: https://a2a-mcp.org/entry/langgraph

Language: en

Category: Frameworks & SDKs

Updated: 2026-09-14

## About

LangGraph is an orchestration framework and runtime for building long-running, stateful agents and controlled agent workflows.

## Key Features

- Represent agent workflows with state, nodes and transitions.
- Build durable execution and human-review steps into agent applications.

## Use Cases

Build application-owned agents with tools, workflow control and model integration.

## Setup or Requirements

For Python, install `pip install -U langgraph` or use `uv add langgraph`. Define the graph state, nodes and edges, and configure the model and tools the workflow calls. Use the separate JavaScript guide for that runtime.

## Compatibility

Python and JavaScript implementations have their own packages and API documentation.

## Limitations

LangGraph supplies orchestration primitives; model access and tool integrations are configured by the application.

## Sources

- Project documentation: https://docs.langchain.com/oss/python/langgraph/overview
- Setup reference: https://docs.langchain.com/oss/javascript/langgraph/graph-api

## Tags

- agent-framework
- stateful-agents
- orchestration

## Alternatives

- Mastra
- Microsoft AutoGen

## Sources

- [Project website](https://docs.langchain.com/oss/python/langgraph/overview)
- [Project documentation](https://docs.langchain.com/oss/python/langgraph/overview)
- [Setup reference](https://docs.langchain.com/oss/javascript/langgraph/graph-api)

Last checked: 2026-09-14