Repolode
Submit
langchain-ai

langchain-ai/langgraph-swarm-py

A Python library for building swarm-style multi-agent systems on LangGraph, where agents hand off control to each other by specialty.

View on GitHubWebsite
Stars
1,589
Forks
228
Open issues
22
Last commit
9 days ago

What is it?

LangGraph Multi-Agent Swarm is a Python library for building multi-agent systems in which agents dynamically hand off the conversation to one another based on their specializations. Each handoff is exposed to the LLM as a tool, so an agent can transfer control to a named peer when a request falls outside its domain.

The swarm tracks which agent was last active and resumes the conversation with that agent on the next turn, provided the graph is compiled with a checkpointer. It is built on LangGraph and supports its streaming, short-term and long-term memory, and human-in-the-loop features. Handoff tools are customizable, and the shared message state can be replaced with per-agent state for agents that should not expose their full history.

Key features

  • 01Creates swarm workflows from a list of agents with a default active agent.
  • 02Provides a prebuilt handoff tool that lets agents transfer control to named peers.
  • 03Remembers the last active agent and resumes with it when a checkpointer is configured.
  • 04Allows custom handoff tools with extra arguments such as a task description.
  • 05Supports per-agent message state to keep agent histories separate from the shared swarm history.

Use cases

  • Building a customer support system where a triage agent routes users to billing or technical specialists.
  • Coordinating math, writing, and research agents that take turns handling a single conversation.
  • Prototyping multi-agent conversations on LangGraph with memory and streaming.

Quick start

  1. 1

    Install the library

    pip install langgraph-swarm
  2. 2

    Install the OpenAI integration and set your API key

    pip install langgraph-swarm langchain-openai
  3. 3

    Export your OpenAI API key and run the Python quickstart example

    export OPENAI_API_KEY=<your_api_key>

Things to know

  • The quickstart example requires a paid OpenAI API key.

In collections

Multi-agent orchestration frameworksOpen-source AI agent frameworks & apps

Topics

#agents#multi-agent#langgraph#llm#python#agent-handoff#llms#multiagent#multiagent-systems

FAQ

What is langgraph-swarm-py?
A Python library for building swarm-style multi-agent systems on LangGraph, where agents hand off control to each other by specialty. LangGraph Multi-Agent Swarm is a Python library for building multi-agent systems in which agents dynamically hand off the conversation to one another based on their specializations. Each handoff is exposed to the LLM as a tool, so an agent can transfer control to a named peer when a request falls outside its domain.
Is langgraph-swarm-py free to use? What license does it have?
langgraph-swarm-py is open source under the MIT license. Check the license terms before using it in commercial products.
Is langgraph-swarm-py still maintained?
Actively maintained. The last commit to langgraph-swarm-py was on Oct 28, 2026.
What can I use langgraph-swarm-py for?
Building a customer support system where a triage agent routes users to billing or technical specialists. Coordinating math, writing, and research agents that take turns handling a single conversation. Prototyping multi-agent conversations on LangGraph with memory and streaming.
How do I install langgraph-swarm-py?
Install the library: pip install langgraph-swarm Install the OpenAI integration and set your API key: pip install langgraph-swarm langchain-openai Export your OpenAI API key and run the Python quickstart example: export OPENAI_API_KEY=<your_api_key>
Where can I get langgraph-swarm-py?
The source code and installation instructions are on GitHub: https://github.com/langchain-ai/langgraph-swarm-py

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

Related projects