data.gouv.fr MCP Server (mcp data gouv)

Overview

The data.gouv.fr MCP Server is the official open-source Model Context Protocol (MCP) server developed by the French government's digital agency (Etalab / data.gouv.fr). It provides a secure, standardized bridge that allows AI agents and chatbots to interact directly with France's national open data platform — home to over 74,000 public datasets — without requiring custom API integrations or API keys.

Launched in early 2026 as an experimental but production-ready service, it uses streamable HTTP transport and exposes structured tools that AI models can call conversationally.

Key Features

• Official & Trusted: Maintained by the data.gouv.fr team; MIT-licensed and fully open source.
• Hosted Endpoint: Ready-to-use at https://mcp.data.gouv.fr/mcp (recommended for most users).
• No API Key Required: Public open data access with no authentication barriers.
• Rich Tool Set: Exposes ~10 structured tools for searching datasets, fetching metadata, downloading resources, exploring organizations, and analyzing content.
• Natural Language Interaction: AI agents can ask questions like “Find recent datasets about French energy consumption” or “Summarize the latest census data for Paris”.
• Python Implementation: Built with the official MCP SDK; easy to self-host if needed.
• SSE / Streamable HTTP Support: Compatible with Claude Desktop, Cursor, VS Code, ChatGPT advanced features, and other MCP clients.

Supported Capabilities

• Search datasets using keywords, themes, organizations, or advanced filters
• Retrieve detailed metadata, licenses, and resource URLs
• Explore producers (organizations) and their published datasets
• Access usage metrics and popularity indicators
• Fetch and analyze specific dataset resources
• Combine multiple tools for complex data exploration workflows

Use Cases

• AI-Powered Open Data Research: Quickly discover and summarize relevant French public datasets for reports, studies, or journalism.
• Data-Driven Decision Making: Let agents analyze government data on economy, environment, health, transport, and more.
• Educational & Civic Tools: Build conversational interfaces that make public data accessible to citizens and students.
• Multi-Agent Automation: Combine with other MCP servers (GitHub, Azure DevOps, weather, etc.) for comprehensive research agents.
• Journalism & Transparency: Enable LLMs to cross-reference official statistics in real time.

Getting Started

1. Use the hosted server (easiest):
   Copy

   {
     "mcpServers": {
       "datagouv": {
         "type": "streamable",
         "url": "https://mcp.data.gouv.fr/mcp"
       }
     }
   }
   

2. Add the configuration in your MCP client (Claude, Cursor, VS Code, etc.).
3. Start chatting — the AI will automatically use the available tools to search and analyze data.gouv.fr datasets.

Self-hosting instructions and full tool documentation are available in the GitHub repository.

Resources

• Official GitHub Repository: https://github.com/datagouv/datagouv-mcp
• data.gouv.fr Announcement: https://www.data.gouv.fr/posts/experimentation-autour-dun-serveur-mcp-pour-datagouv
• Hosted MCP Endpoint: https://mcp.data.gouv.fr/mcp

This server represents France's leadership in making high-quality public data natively accessible to the AI ecosystem through the open Model Context Protocol.