Mapbox

Converts human-readable addresses and place names to geographic coordinates (latitude/longitude) using the Mapbox Geocoding V6 API. Implements schema-based input validation via Zod to normalize address strings, handles authentication through MAPBOX_ACCESS_TOKEN environment variable, and returns structured location data with confidence scores and bounding boxes for spatial disambiguation.

Converts geographic coordinates (latitude/longitude) back into human-readable addresses and location context using Mapbox Geocoding V6 API. Accepts coordinate pairs as input, validates them through Zod schemas, and returns hierarchical location information (street address, city, region, country) with proximity-based ranking for ambiguous locations.

Provides pre-built integration configurations for popular AI clients: Claude Desktop (via claude_desktop_config.json), VS Code (via extension), and Smolagents (Python framework). Each integration handles MCP server discovery, tool registration, and client-specific configuration. Enables AI agents in these environments to invoke Mapbox geospatial tools without manual setup.

Calculates optimal routes between two or more points supporting multiple transportation modes (driving, walking, cycling) with real-time traffic awareness. Uses Mapbox Directions API to compute turn-by-turn instructions, distance, duration, and geometry. Implements mode-specific routing logic and traffic-aware duration estimates through the MapboxApiBasedTool pattern with Zod schema validation for waypoints and routing parameters.

Calculates efficient one-to-many, many-to-one, or many-to-many travel time and distance matrices between multiple origin and destination points using Mapbox Matrix API. Optimized for bulk distance/duration lookups without computing full route geometry, returning a matrix of travel times and distances. Implements coordinate validation and matrix parameter handling through MapboxApiBasedTool base class.

Generates isochrone polygons representing areas reachable from a point within specified time or distance constraints using Mapbox Isochrone API. Computes accessibility zones for different transportation modes and returns GeoJSON polygons that can be visualized or analyzed. Implements time/distance parameter validation and polygon generation through MapboxApiBasedTool pattern.

Discovers specific points of interest (POIs) by name or brand within a geographic area using Mapbox Search API. Accepts search queries and optional proximity coordinates, returns ranked results with location data, categories, and metadata. Implements query normalization and proximity-based ranking through MapboxApiBasedTool with Zod schema validation for search parameters.

Discovers points of interest by category (restaurants, hotels, gas stations, parks, etc.) within a geographic area using Mapbox Search API category filtering. Accepts category names or codes and optional proximity/bounding box constraints, returns ranked results filtered by POI type. Implements category validation and spatial filtering through MapboxApiBasedTool pattern.

Generates customizable static map images with markers, overlays, and various map styles using Mapbox Static Images API. Accepts map configuration (center, zoom, style, markers, polygons) and returns PNG/JPEG images suitable for embedding in documents or chat interfaces. Implements style selection, marker positioning, and overlay rendering through MapboxApiBasedTool with Zod schema validation for image parameters.

Implements Model Context Protocol (MCP) server that adapts Mapbox REST APIs into standardized tool interface for AI clients (Claude Desktop, VS Code, Smolagents). Handles unified authentication via MAPBOX_ACCESS_TOKEN environment variable, tool registration, request routing, and response formatting. Uses abstract MapboxApiBasedTool base class implementing template method pattern for consistent error handling and API communication across all geospatial tools.

Implements runtime input validation using Zod schemas for all tool parameters, ensuring type safety and consistent error messages across all geospatial operations. Each tool defines inputSchema property with Zod validators for coordinates, addresses, modes, and other parameters. Validation occurs before API invocation, catching malformed inputs and providing structured error responses to AI clients.