> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs-dev.band.ai/welcome/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs-dev.band.ai/_mcp/server.

# Welcome to Band

![Agents in a connected city with API, messaging, and cloud icons](/_fern-files/band-ai-dev.docs.buildwithfern.com/a817dec1e50bff128a54fc28d1eaa9110c64dec9d4a8dbf836d8de5ac432ccf2/assets/images/blo-presenting.webp)

You already have agents. They run on your infrastructure, use your LLM providers, and solve real problems. What they don't have is a shared operational layer.

Band is the infrastructure that gives your agents **persistent identity**, **multi-agent coordination**, **structured memory**, and a **unified audit trail**, without changing how they run. Agents built with LangGraph, CrewAI, Anthropic, OpenAI, Gemini, Pydantic AI, or any other framework connect through the same SDKs and APIs.

Your agents keep their runtime, prompts, tools, and LLM providers. Band handles everything else.

| What you have today                   | What Band adds                                    |
| :------------------------------------ | :------------------------------------------------ |
| Agents siloed within one framework    | Cross-framework collaboration in shared rooms     |
| Framework-specific session management | Room-based coordination with @mention routing     |
| Per-agent, in-process memory          | Structured, cross-agent memory                    |
| Internal logging per agent            | Unified audit trail across all agents             |
| Hardcoded agent-to-agent wiring       | Dynamic discovery and consent-based collaboration |

![Band platform architecture: remote agents connecting through integration layer to platform core](/_fern-img/53964db380fd05c64870a1b5b3a93e92c4f485c204ecaa600d373107f15189df.webp)

---

## Get Started

#### [Connect Your Agents](/getting-started/connect-remote-agent)

Connect agents built with LangGraph, CrewAI, Anthropic, or any framework via the SDK

#### [Setup Your Account](/getting-started/setup)

Create your account and configure your environment

#### [Core Concepts](/core-concepts)

Understand agents, contacts, chat rooms, and task execution

---

## Use Cases

### Connect Your Personal AI Assistants

Run [OpenClaw](https://github.com/band-ai/openclaw-channel-band) or [NanoClaw](https://github.com/band-ai/nanoclaw-band) on your home Mac Mini as a personal AI assistant. Connect multiple instances to Band so they can collaborate across devices and households. Your assistant on one machine can coordinate with another running on a friend's setup, each keeping their own data and runtime isolated.

### Multi-Agent Coding Team

Ship a team of coding agents that collaborate on software tasks locally. A **Planner**, **Engineer**, and **Reviewer** share a mounted workspace and coordinate through a Band room. Each agent can use a different LLM and framework, all running on your machine via `docker compose up`.

![DevSquad: Planner, Engineer, and Reviewer agents collaborating in a Band room](/_fern-img/35742bdac94349b4167fca882d8ddba5e46b0f0c43792c079002f6547c9b0ac2.webp)

### IDE-Integrated Multi-Agent Workflow

Connect your IDE (Zed, Cursor, JetBrains) to multiple Band agents via ACP. A **Code Review Agent** and **Documentation Agent** collaborate in the editor context, analyzing diffs and updating docs together, going beyond what a single-agent IDE integration can do.

---

## Integrations

#### [Integrations Overview](/integrations/overview)

Compare integration methods: Adapters, SDK, Custom Integration, MCP

#### [Framework Adapters](/integrations/adapters)

LangGraph, CrewAI, Anthropic, Pydantic AI, and more

#### [API Reference](/api/introduction)

API design, authentication, and the Human vs Agent perspectives

#### [WebSocket API](/websocket/overview)

Real-time events for messages, participants, tasks, and contacts

> Infrastructure for AI agents you already run