Mapbox MCP. Calculate real-world travel metrics for any location.

This Mapbox server gives your agent world-class location services. You run complex spatial calculations—from mapping out optimal delivery routes to figuring out service areas—all through natural conversation. It's the core toolset you need when your app relies on solid, reliable geospatial data.

Geocoding and Addressing
When you pass a place name or full address, the geocode tool spits back precise latitude/longitude pairs along with detailed components like the city, postal code, and structured address parts. If you grab raw coordinates, the reverse_geocode tool translates them right back into the nearest readable street address and location details.

The system also lets you run search_nearby to pinpoint specific points of interest—say, a 'cafe' or a 'hospital'—near given coordinates; it returns their names and distances.

Routing and Movement
Calculating paths is what this server does best. You can use get_directions to generate step-by-turn instructions for driving, walking, or cycling between two points; the result includes the total distance traveled and the estimated duration. For bigger logistics jobs, you'll want to hit up get_distance_matrix. This tool takes multiple origin and destination points and spits out a matrix showing the travel time (in seconds) and distance (in meters) for every possible pairing, which is crucial for fleet management.

Determining how high up something is? Run get_elevation on any set of coordinates to get elevation data, helpful if you're analyzing terrain changes.

Service Areas and Mapping Output
Need to define a service boundary? Use get_isochrone. This tool draws polygon contours that map out every single area your client can reach from an origin point within a specific time or distance. You can also generate direct visual assets with get_static_map, which gives you an image URL for any location, letting you embed the map directly into markdown or HTML reports.

How It Works For Your Agent
It's straightforward. Your AI client uses this server by calling specific functions like geocode or get_directions. You don't write code; you just tell your agent what you want—like, 'What's the fastest route from 123 Main Street to the airport?' The agent handles passing that data through the correct tool.

It returns structured data you can immediately use in your application logic.

For example, if you need to know all potential delivery spots within a five-mile radius of a warehouse, you'll combine get_isochrone with search_nearby. If your client needs to validate an old GPS ping and figure out what address it corresponds to, that's the reverse_geocode tool for. The system handles all these complex spatial relationships—from pinpointing addresses to calculating travel time matrices—so you don't have to manage multiple mapping APIs or deal with messy data structures yourself.