MathModDB MCP Server

MathModDB MCP server for semantic ontology exploration and SPARQL querying.

What it provides

• Explore_Ontology: semantic schema discovery (classes/properties/qualifiers).
• SPARQL_Query: single SPARQL query execution.
• Batch_SPARQL_Query: batch SPARQL query execution.

The server runs as streamable-http on port 8000 by default.

See the MCP server in action in this example chat

Requirements

• Python 3.12+ (for local runs)
• uv (for local runs)
• OPENAI_API_KEY (required)
• Docker (for containerized runs)
• Docker BuildKit enabled (required by the Dockerfile build secret)

Run locally

uv sync --frozen
export OPENAI_API_KEY=your_key_here
uv run fastmcp run main.py --host 0.0.0.0 --port 8000 --transport streamable-http

Build and run with Docker

The Docker image initializes ontology/vector data during build, so the API key must be provided at build time as a Docker secret.

Build requirements

• Docker BuildKit enabled (DOCKER_BUILDKIT=1)
• OPENAI_API_KEY present in your shell environment

Build image

DOCKER_BUILDKIT=1 docker build \
  --secret id=OPENAI_API_KEY,env=OPENAI_API_KEY \
  -t mathmoddb-mcp .

Run container

docker run --rm --name mathmoddb-mcp -p 8000:8000 -e OPENAI_API_KEY="$OPENAI_API_KEY" mathmoddb-mcp

You can also use the included helper script:

./start.sh

Use with Claude Desktop

After the server is running at http://127.0.0.1:8000/mcp, add this MCP entry in your Claude Desktop config:

"MathModDB": {
  "command": "npx",
  "args": [
    "mcp-remote",
    "http://127.0.0.1:8000/mcp"
  ],
  "env": {}
}

Claude Desktop-side prerequisites

• Node.js / npm available locally (for npx)
• mcp-remote runnable via npx