mcp-server-agribalyse

What is mcp-server-agribalyse?

mcp-server-agribalyse is a Model Context Protocol (MCP) server that enables Large Language Models to query the ADEME Agribalyse 3.1 dataset. It provides tools, resources, and prompts for retrieving and analyzing environmental impact data (climate impact, water use, ecotoxicity, etc.) on thousands of food products.

How to use mcp-server-agribalyse?

You can inspect server behavior with npx @modelcontextprotocol/inspector uvx run src/agribalyse/server.py or follow logs with mcp dev server.py. To test interactions, run the example client via python client/client.py. Use pytest tests/test_mcp_tools to execute the unit tests.

Key features of mcp-server-agribalyse

• Eight tools for querying rows, aggregating metrics, and retrieving schema.
• Seven resource URIs for exploring dataset fields, files, and columns.
• Seven predefined prompts for product search, comparison, and explanation.
• Based on the FastMCP framework with active maintenance.
• Public API integration – no local database required.
• Full OpenAPI specification available via read_api_docs tool.

Use cases of mcp-server-agribalyse

• Ask an LLM for the carbon footprint of a specific food product.
• Compare two food products by a chosen environmental indicator.
• Retrieve average or percentile values for numeric fields across a filtered subset.
• Explore the dataset schema and available filter values before building complex queries.

FAQ from mcp-server-agribalyse

What is the Agribalyse dataset?

Agribalyse is a dataset published by ADEME that contains environmental indicators (e.g., climate impact, water use, ecotoxicity) for thousands of food products.

How can I debug or inspect the server?

Use the MCP inspector: npx @modelcontextprotocol/inspector uvx run src/agribalyse/server.py, or run mcp dev server.py to follow logs.

How do I run the tests?

Activate your virtual environment and run pytest tests/test_mcp_tools to execute the unit test suite.

Who maintains this server?

The server is maintained by Tracy André, affiliated with Positive Solutions. Contact: [email protected].

Does the server require a local database or API key?

No – the server queries the public ADEME Agribalyse API directly. No local database or API key is mentioned in the README.