> ## Documentation Index
> Fetch the complete documentation index at: https://mintlify.com/ably/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# LangGraph integration

Integrate Ably AI Transport with LangGraph to build reliable, resumable AI agent experiences.

## Overview

LangGraph is a framework for building stateful, multi-agent applications with LLMs. By combining LangGraph with Ably AI Transport, you can:

* Stream agent responses reliably to users across devices
* Enable users to resume agent interactions after disconnections
* Support multi-user collaborative agent sessions
* Implement durable execution patterns with session persistence

## Getting started guides

Explore the following guides to integrate LangGraph with AI Transport:

<Tiles>
  {[
    {
      title: 'Message-per-response',
      description: 'Stream LangGraph responses using message appends',
      image: 'icon-tech-javascript',
      link: '/docs/guides/ai-transport/langgraph/lang-graph-message-per-response',
    },
    {
      title: 'Message-per-token',
      description: 'Stream LangGraph responses using individual token messages',
      image: 'icon-tech-javascript',
      link: '/docs/guides/ai-transport/langgraph/lang-graph-message-per-token',
    },
    ]}
</Tiles>

## Key features

### Streaming responses

Stream LangGraph agent responses in realtime using either the [message-per-response](/docs/ai-transport/token-streaming/message-per-response) or [message-per-token](/docs/ai-transport/token-streaming/message-per-token) pattern.

### Agent resumption

LangGraph provides durable execution for agents. Combine this with Ably's session management to enable agents to resume after restarts. Learn more about [resuming sessions](/docs/ai-transport/sessions/resuming-sessions).

### Multi-agent coordination

Use Ably channels to coordinate multiple LangGraph agents in collaborative workflows.

## Next steps

* Follow the [message-per-response guide](/docs/guides/ai-transport/langgraph/lang-graph-message-per-response) to get started
* Learn about [token streaming patterns](/docs/ai-transport/token-streaming/overview)
* Explore [agent resumption](/docs/ai-transport/sessions/resuming-sessions)