cablate/mcp-google-map

What it solves

It eliminates the need for developers to manually rebuild complex location-based workflows—such as place searching, routing, and itinerary planning—for every AI agent. Instead of giving an agent raw API access, it provides a structured set of tools and "Skills" that handle the multi-step logic required to get useful real-world location outcomes.

How it works

The project acts as a bridge between AI agents and the Google Maps Platform. It converts Google Maps capabilities into 18 read-only tools and three specialized Skills (General Research, Travel Planning, and Local SEO). It can be integrated in several ways:

• MCP (Model Context Protocol): Connects to clients like Claude Desktop, Cursor, or VS Code via stdio or Streamable HTTP.
• Plugins: Integrates directly with Codex or Claude Code without requiring a full MCP setup.
• Standalone CLI: Provides a stateless command-line interface for scripts and automation.

Who it’s for

• AI Agent Developers: Those building apps that need reliable location data, routing, and geographic planning.
• Travel & Logistics Planners: Users creating agents to build itineraries or optimize delivery routes.
• Local SEO Specialists: Those using AI to audit business visibility and track geographic rankings.

Highlights

• Outcome-Oriented Tools: Includes composite tools for comparing places, optimizing multi-stop routes (up to 25 stops), and exploring neighborhoods.
• Flexible Integration: Supports MCP stdio, remote HTTP servers, and direct plugin installations for Codex and Claude Code.
• Comprehensive Data Access: Provides tools for geocoding, distance matrices, weather, air quality, and static maps.
• Security-First Credentials: Supports API key isolation via request headers for multi-tenant HTTP deployments.