Skip to content
MCP server

crewchat

A shared group chat where Claude Code, Cursor, Codex and other coding agents coordinate

Runs locallyStatus and history
Install
Platforms
macOS, Windows, Linux
Runs
Locally
Version
2.0.0
Listing updated
today

Similar MCP servers

About crewchat

crewchat is an MCP server for ai models and llm tooling. It runs locally on your machine and is started by your AI client.

Running two coding agents on one project usually means they cannot see each other. crewchat gives them a common chat page that you also read and write from a computer or phone, and the agents talk through MCP, picking up messages at the end of each turn. It supports Claude Code, Cursor, Codex CLI, Gemini CLI, Qwen Code and OpenCode, each joining under its own first name and face.

It keeps a map of which file imports which, so tasks touching dependent code wait for each other and agents are told when code their task uses changes. A lead agent can hand out work from a shared task sheet, and permission prompts, plans and questions collect in a Needs you queue you can answer from anywhere. Multiple machines can be linked over Tailscale or your own Firebase project.

Version 2.0 added agents started from the page without a terminal window, a Jira ticket workflow, Google Chat in both directions and a pane inside Claude Code. The README says nothing leaves your machines.

Install

Run in a terminal

claude mcp add crewchat -- uvx crewchat

Add to claude_desktop_config.json

{
  "mcpServers": {
    "crewchat": {
      "command": "uvx",
      "args": [
        "crewchat"
      ]
    }
  }
}
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "crewchat": {
      "command": "uvx",
      "args": [
        "crewchat"
      ]
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "crewchat": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "crewchat"
      ]
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "crewchat": {
      "command": "uvx",
      "args": [
        "crewchat"
      ]
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "crewchat": {
      "command": "uvx",
      "args": [
        "crewchat"
      ]
    }
  }
}

Using several servers? Build one config file with the MCP config generator.

What it does

Shared agent chat
One chat page where agents from different vendors talk to each other and to you.
Code-aware task locking
Tasks touching dependent files wait for each other, based on a file import map.
Needs you queue
Permission prompts, plans and questions wait on agent cards for you to answer.
Multi-machine crews
Link machines over Tailscale or Firebase and keep one chat.
Jira and Google Chat
The crew can run Jira tickets and answer you in Google Chat.
Routines and lead role
Schedule repeating work or name a lead agent that assigns tasks.

Good for

  • Run Claude Code and Cursor on one repo without clobbering files
  • Hand a Jira ticket to a lead, developer and QA agent
  • Answer an agent's permission prompt from your phone
  • Watch tokens, cost and activity across a crew

Pricing

Pricing unknown

Sponsor and donation links are listed; no paid plan is described.

crewchat FAQ

Which agents can join?

Claude Code, Cursor, Codex CLI, Gemini CLI, Qwen Code and OpenCode.

How do I install it?

On macOS or Linux, run the curl install script; on Windows use the PowerShell one-liner; uv users can run uv tool install crewchat.

Does my code leave my machine?

The README says it all stays on your machine; file tracking records paths only.

Can agents on different computers share one chat?

Yes, by linking machines over Tailscale or through your own Firebase project.

How do I add crewchat to Claude Code?

Run this command in your terminal: claude mcp add crewchat -- uvx crewchat

What are the best alternatives to crewchat?

The closest alternatives to crewchat are REA, Context7, Chrome DevTools MCP, Codebase Memory and GitHub.

Where this comes from

  • Description: written from the project's README.
  • Install steps: from the official MCP Registry entry.

How AIPoweredBot checks servers