TransformerBee

You have a Python-based MCP server that lets Claude Desktop and other MCP clients talk to TransformerBee. It can be run locally from a Python virtual environment or started from a Docker container, so you can choose the setup that fits your workflow. This server exposes the TransformerBee MCP interface and uses standard MCP tooling to connect, request data, and trigger actions between your client and transformer.bee.

How to use

Start the MCP server from your preferred environment and connect your MCP client (for example, Claude Desktop) to the server. You can run it directly in a Python virtual environment or launch it via Docker. When running locally, set TRANSFORMERBEE_HOST to the address where transformer.bee is reachable (for example, http://localhost:5021). Then configure your MCP client with this server so it can request data and perform actions through the TransformerBee MCP interface.

How to install

Prerequisites: you need Python and either a Python package manager or Docker to get started.

Install as a Python package using uv or pip:

uv install transformerbeemcp

pip install transformerbeemcp

Install as a Docker image:

docker pull ghcr.io/hochfrequenz/transformerbee.mcp:latest

Start the server via CLI from within the virtual environment where you installed the package:

(myvenv) run-transformerbee-mcp-server

Or start the server using Docker with the required environment variable for the host where transformer.bee is exposed:

docker run --network host -i --rm -e TRANSFORMERBEE_HOST=http://localhost:5021 ghcr.io/hochfrequenz/transformerbee.mcp:latest

Additional configuration and notes

Configure Claude Desktop to register this MCP server. If you have checked out the transformerbee MCP code path, you can install the server entry in Claude Desktop with the following entry in the mcpServers section:

{
  "mcpServers": {
    "TransformerBee.mcp": {
      "command": "C:\\github\\MyProject\\.myvenv\\Scripts\\run-transformerbee-mcp-server.exe",
      "args": [],
      "env": {
        "TRANSFORMERBEE_HOST": "http://localhost:5021",
        "TRANSFORMERBEE_CLIENT_ID": "",
        "TRANSFORMERBEE_CLIENT_SECRET": ""
      }
    }
  }
}

If you install the package via pip or uv, you can modify Claude Desktop settings to point to your local start command. The example below shows how you would reference the virtual environment path and the host URL used by transformer.bee. If you have not configured the handy CLI command, you can invoke Python with non-empty arguments to start the server.

Note that this package marks uv as a dev-dependency, so you might need to install it with the extra flag: pip install transformerbeempc[dev] in your virtual environment as well as enabling MCP tooling. For details about the environment variables and starting transformer.bee locally, refer to the transformerbeeclient.py docs.

If you prefer Docker, you can configure Claude Desktop to start the server with a configuration like this, which mirrors the Docker command example shown above:

{
  "mcpServers": {
    "TransformerBee.mcp": {
      "command": "docker",
      "args": [
        "run",
        "--network",
        "host",
        "-i",
        "--rm",
        "-e",
        "TRANSFORMERBEE_HOST=http://localhost:5021",
        "ghcr.io/hochfrequenz/transformerbee.mcp:latest"
      ],
      "env": {
        "TRANSFORMERBEE_HOST": "http://localhost:5021",
        "TRANSFORMERBEE_CLIENT_ID": "",
        "TRANSFORMERBEE_CLIENT_SECRET": ""
      }
    }
  }
}