SPARQL LLM

The easiest approach is to point your MCP client at the publicly hosted Streamable HTTP endpoint. No API key, no local install — just add the URL to your client configuration.

{
  "servers": {
    "expasy-mcp-http": {
      "url": "https://chat.expasy.org/mcp/",
      "type": "http"
    }
  }
}

For local stdio deployment — for example when you want to point at your own SPARQL endpoints — install sparql-llm via pip or uv.

pip install sparql-llm
# or with uv:
uv add sparql-llm

If you are using your own SPARQL endpoints, create a JSON settings file that lists them. The SETTINGS_FILEPATH environment variable tells the server where to find this file.

{
  "app_org": "My Organization",
  "app_topics": "proteins, genes, and chemical compounds",
  "endpoints": [
    {
      "label": "UniProt",
      "endpoint_url": "https://sparql.uniprot.org/sparql/",
      "description": "UniProt protein sequence and annotation data."
    }
  ]
}

Add the sparql-llm server to your client configuration using uvx as the command. Set SETTINGS_FILEPATH to the absolute path of the JSON settings file you created.

{
  "servers": {
    "expasy-mcp": {
      "type": "stdio",
      "command": "uvx",
      "args": ["sparql-llm"],
      "env": {
        "SETTINGS_FILEPATH": "/Users/you/sparql-mcp.json"
      }
    }
  }
}

Ask your AI client to list available MCP tools. You should see tools for retrieving relevant documents, retrieving class schemas, and executing SPARQL queries against configured endpoints.