Galaxy MCP

Galaxy MCP is a Model Context Protocol server that connects any MCP-capable AI assistant — Claude Desktop, Orbit/Loom, and others — to a Galaxy bioinformatics instance. Point it at any Galaxy server with a URL and API key, and the agent can search and run tools, drive workflows, and manage histories and datasets through Galaxy’s API (via BioBlend).

It is the bridge that lets a conversational agent do real, reproducible compute on Galaxy instead of just talking about it.

Capabilities

Connect to any Galaxy instance (URL + API key), with optional browser-based OAuth that mints short-lived keys per session.
Tools — search the catalog, view tool details, execute tools.
Workflows — access and import from the Interactive Workflow Composer (IWC).
Histories & datasets — create, inspect, and manage.
File upload from local storage into Galaxy.
Ships stdio (local) and streamable-HTTP (remote / multi-user) transports.

Quick start

Run the stdio server directly with uv:

uvx galaxy-mcp

It reads GALAXY_URL and GALAXY_API_KEY from the environment. You can also install it from PyPI with pip install galaxy-mcp.

Connect Claude Desktop

Add Galaxy MCP to claude_desktop_config.json:

{
  "mcpServers": {
    "galaxy-mcp": {
      "command": "uvx",
      "args": ["galaxy-mcp"],
      "env": {
        "GALAXY_URL": "https://usegalaxy.org",
        "GALAXY_API_KEY": "your-api-key"
      }
    }
  }
}

Restart Claude Desktop and the Galaxy tools become available in the conversation.

With Orbit and Loom

Orbit and the Loom CLI use Galaxy MCP under the hood — when you connect Galaxy in Orbit, or set GALAXY_URL / GALAXY_API_KEY for the CLI, the agent routes work through uvx galaxy-mcp automatically. See Getting started.

The source lives at galaxyproject/galaxy-mcp.