MCP Datapi

🚀 Climate Data Store (CDS) MCP Server

This is an implementation of a Model Context Protocol (MCP) server that provides an interface for large language models to access CDS catalog data and task status. The underlying API is datapi, and its documentation can be found here.

🚀 Quick Start

Prerequisites

• Python 3.13 or higher.
• CDS API key: Obtain it here.
• MCP host/client: Tested on Claude Desktop and MCP Monitor.

Installation

• Clone the repository:

git clone git@github.com:albertdow/mcp-datapi.git
cd mcp-datapi

• Install dependencies (using uv):

uv add "mcp[cli]" datapi python-dotenv

• Set the CDS API key by creating a .env file and adding the following content:

DATAPI_URL=<DATAPI_URL>
DATAPI_KEY=<DATAPI_KEY>

For detailed information on setting the CDS API key, see here.

Usage

Development Mode with MCP Monitor

Test the server locally:

mcp dev datapi_server.py

Integration with Claude Desktop

mcp install datapi_server.py --name "DatapiServer" -f .env

Or add it directly to claude_desktop_config.json:

{
    "mcpServers": {
        "DatapiServer": {
            "command": "uv",
            "args": [
                "--directory",
                "mcp-datapi",
                "run",
                "datapi_server.py"
            ],
            "env": {
                "DATAPI_URL": "<DATAPI_URL>",
                "DATAPI_KEY": "<DATAPI_KEY>"
            }
        }
    }
}

⚠️ Important Note

You need to specify the path of uv, for example, /Users/username/.local/bin/uv.

✨ Features

Tools

• get_jobs: Find available jobs, optionally add filters based on status. Returns a list of job IDs.
• download_job_result: Download job results using the job ID.
• get_all_collections: Get the IDs of all available collections in the catalog.
• get_collection_by_id: Get relevant information by specifying the collection ID.
• submit_job: Submit a download request.

Environment Variable Support

It supports using the .env file.