Opendata Ademe

@data-fair/mcp

A Model Context Protocol (MCP) server to interact with the Data Fair ecosystem.

Usage modes

This MCP server can be used in two different modes:

Standalone mode (Docker)

Run the MCP server as a standalone Docker container that connects remotely to a Data Fair instance. This mode uses the stdio transport and is suited for integrating with LLM clients (e.g., Claude Desktop, VS Code).

• PORTAL_URL is required — set it to the base URL of the Data Fair portal you want to query.
• DATA_FAIR_API_KEY — optional API key for authenticating requests to the Data Fair instance.

{
  "mcpServers": {
    "data-fair-datasets": {
      "command": "docker",
      "args": ["run", "-i", "--rm", "-e", "PORTAL_URL=https://opendata.koumoul.com", "ghcr.io/data-fair/mcp"]
    }
  }
}

Data Fair stack mode (web service)

Deploy the MCP server as a web service alongside Data Fair in the same infrastructure stack. In this mode, the server is typically available on /mcp-server on the same domain as the Data Fair instance (Data Fair is deployed on /data-fair, a portal on /, etc.).

Example endpoint: https://opendata.koumoul.com/mcp-server/datasets/mcp

• Uses the http transport (TRANSPORT=http).
• PORTAL_URL is optional — it is automatically inferred from reverse-proxy headers (X-Forwarded-Host, X-Forwarded-Proto, X-Forwarded-Port).
• IGNORE_RATE_LIMITING — optional secret to bypass Data Fair rate limiting constraints.
• Authentication is handled by the reverse proxy and session management of the Data Fair stack.

⚙️ Environment Variables

🔨 Development

Take a look at the contribution guidelines.