Skip to content

Installation

There are three ways to run the TfL MCP server: Node.js MCPB (recommended), Docker, or Java direct.

Both the Node.js and Java implementations expose the same 12 tools with identical behaviour — choose whichever fits your environment.

Claude Desktop Configuration

Open your Claude Desktop configuration file. Usually, this is located at: - macOS: ~/Library/Application Support/Claude/claude_desktop_config.json - Windows: %APPDATA%\Claude\claude_desktop_config.json - Linux: ~/.config/Claude/claude_desktop_config.json

Add the following to your mcpServers object, choosing one of the options below:

The lightest option — no Docker or Java required. Just Node.js 20+.

Download tfl.mcpb from GitHub Releases and install it in Claude Desktop, or add manually:

{
  "mcpServers": {
    "tfl": {
      "command": "npx",
      "args": ["-y", "@oneill9/mcp-server-tfl"],
      "env": {
        "TFL_APP_KEY": "your_api_key_here"
      }
    }
  }
}

Option 2: Docker

You don't need Java or Node.js installed. Docker will download and run the container securely.

{
  "mcpServers": {
    "tfl": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "TFL_APP_KEY=your_api_key_here",
        "ghcr.io/oneill9/mcp-server-tfl:latest"
      ]
    }
  }
}

Option 3: Java Direct

If you prefer to run it using the built distribution ZIP attached to GitHub Releases:

  1. Download the latest mcp-server-tfl.zip from GitHub Releases.
  2. Unzip it somewhere on your machine.
  3. Update your claude_desktop_config.json:
{
  "mcpServers": {
    "tfl": {
      "command": "/path/to/extracted/mcp-server-tfl/bin/mcp-server-tfl",
      "args": [],
      "env": {
        "TFL_APP_KEY": "your_api_key_here"
      }
    }
  }
}

Obtaining the TfL API Key

See the API Keys page for details on how to generate the TFL_APP_KEY.