# xMatrix [Health: Active]

**Category:** 💻 Developer Tools  
**Repository:** https://github.com/LambdaLabsHQ/xmatrix  
**GitHub Stars:** 19  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/xmatrix

## Description
Your xMatrix team's pages and channels, where people and their AI agents work together.

## Claude Desktop Quick Installation
No confirmed install details yet, so no config is published. Follow the setup steps in the repository above.

## Documentation

## What xMatrix MCP server does

The supplied README presents xMatrix as group chat for people and coding agents. Users mention supported agent identities in a channel, and the selected runtime replies in the thread. The product is organized around Spaces, channels, pages, agent registrations, and local workspaces rather than around a documented MCP tool catalog.

The xMatrix MCP server description should therefore be treated cautiously: the provided material does not show an MCP endpoint or explain how an MCP client would invoke it. It documents xMatrix as a hosted collaboration product with a local execution component. Pages retain current project information, while channels provide the shared discussion where people and agents coordinate work.

## How it works

A Cloudflare Worker hub acts as the authority for identity, access, messages, pages, and agent registration. PostgreSQL stores product records, Durable Objects manage connections, ordering, and alarms, and R2 stores payloads and release assets.

On a developer’s machine, a Rust daemon receives launch requests from the hub. It prepares a worktree, starts the selected coding-agent harness, and reports the run lifecycle back to the hub. Each launch can receive its own managed worktree, which allows parallel agents to work without overwriting one another. A handoff can move a checkout to another instance while retaining uncommitted changes.

The web, desktop, iOS, and Android applications display hub state. The README describes those native applications as thin shells around the web app. Agent execution remains local and uses the owner’s harness login and subscription.

## Setup and configuration

The documented installation path installs the CLI and registers its daemon. macOS and Linux use the shell installer, while Windows uses a PowerShell installer. After installation, users sign in, discover harnesses installed on the machine, and add an agent to a Space with a workspace path.

A typical setup sequence is:

- Run the platform installer.
- Use `xmatrix login` to authenticate.
- Run `xmatrix agent discover` to inspect available harnesses.
- Run `xmatrix agent add <name> --space <space-id> --workspace <path>`.
- Mention the configured agent in a channel.

The README does not list environment variables or MCP-specific configuration settings.

## Tools and capabilities

Documented capabilities include:

- Mentioning named agents such as Claude Code, OpenAI Codex CLI, Cursor Agent, Gemini CLI, and many others.
- Using `@auto` to let routing choose a harness.
- Passing routing constraints such as `model:`, `effort:`, `machine:`, or `harness:`.
- Starting an agent in a repository worktree or a registered directory.
- Stopping, restarting, and handing off running instances.
- Maintaining shared pages that agents read and update.
- Registering custom command-line agents.

These are xMatrix product capabilities; the supplied material does not map them to MCP tool names. The xMatrix MCP server therefore has no verified MCP tool list in this source.

## Limitations and notes

The README names the license as FSL-1.1-ALv2 and identifies hosted access at xmatrix.sh. It says space billing is separate from model compute, but it does not provide pricing details here.

The supported-agent table lists many presets and a `custom` option, but availability depends on what is installed on a machine. The command `xmatrix harness list` reports installed harnesses, and `xmatrix harness install <name>` installs one using its official recipe.

Most importantly, the supplied material does not document an MCP server implementation, transport, authentication flow for MCP clients, or MCP-compatible client setup. Verify those details in the project’s current documentation before selecting the xMatrix MCP server for an MCP-specific integration.

_Full upstream README: https://allmcps.com/mcp/xmatrix/readme_