Routing questions cost Google API calls whoever asks them. Running the server yourself keeps the key, the quota and the per-IP limit under your control, and the four tools stay on the practical side of Maps: which way, how long, how much.
A Node MCP server for Google Maps, packaged for Google Cloud Run and reached over SSE. Four tools covering routing, route comparison, live traffic and trip cost, with rate limiting built in.
- An optimal driving route calculated against real-time traffic
- Several routing alternatives compared under different options
- Current traffic conditions and travel-time analysis
- Trip cost estimated including fuel and tolls
- Your own deployed service, so the endpoint, the key and the quota stay yours
- Per-IP rate limiting, counted by the hour
A GCP account with billing enabled, the gcloud CLI, and Node.js 18+ for local work. Enable the Directions and Geocoding APIs (Maps JavaScript is optional, for polyline encoding), put a Google Maps API key in the environment, then deploy with ./simple-deploy.sh <project-id> <region> and give your client the service URL with /sse on the end. The deployed service takes no authentication of its own — it is public, and rate-limited per IP by the hour.
One command plus a key — npx -y mcp-remote https://YOUR-SERVICE-URL.run.app/sse, then supply credentials