TartuNLP

You can access TartuNLP translation services through an MCP server that lets AI assistants translate text across 700+ language pairs. It supports specialized European and Finno-Ugric languages and is easy to run with Docker or locally for development and testing.

How to use

Use an MCP client to connect to the TartuNLP MCP server and call its tools to translate text or to retrieve the list of supported languages. The server exposes translate_text to translate content between any supported language pair and get_supported_languages to fetch the complete set of available pairs and translation models. Start the server with your preferred method (Docker container or Python), then point your MCP client to the corresponding command or container URL.

How to install

Prerequisites you need before installation: Docker (for the Docker-based deployment) and Python 3.8 or higher (for local development). Ensure you have a working internet connection to pull images or install dependencies.

Choose one of the two installation methods below and follow the steps precisely.

{
  "mcpServers": {
    "TartuNLP": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "ghcr.io/tbitu/tartunlp-mcp:latest"
      ]
    }
  }
}

Additional steps and notes

If you prefer local development instead of Docker, install the package and run the MCP server module directly. Two command variations are shown for Linux/macOS and Windows.

Linux/macOS start command for local development:

{
  "mcpServers": {
    "TartuNLP": {
      "command": "python3",
      "args": ["-m", "tartunlp_mcp_server"]
    }
  }
}

Security and maintenance

Keep the Docker image up to date by pulling the latest tag regularly. For local development, ensure Python dependencies are installed with the -e editable install to reflect source changes quickly.

Troubleshooting

If the server fails to start, verify that the command and arguments are correct for your environment. On Linux/macOS, use python3 to run the module; on Windows, use python. Ensure version requirements (Python 3.8+) are met and that Docker can access the container registry.

Available tools

translate_text

Translates text between any supported language pair, enabling multilingual communication and content localization.

get_supported_languages

Retrieves the full list of supported language pairs and available translation models from the MCP server.