Púca MCP Server

The Osm‑UI‑Puca MCP server bridges the gap between AI assistants and OpenStreetMap data by exposing a rich set of location‑centric tools over the Model Context Protocol. It solves a common pain point for developers: obtaining precise geographic information, such as coordinates or nearby amenities, without having to write custom API wrappers or manage authentication for multiple map services. By encapsulating these queries behind a single MCP endpoint, Púca allows an AI assistant to ask for “the nearest defibrillator” or “a fast‑food place within 500 m of this address,” and receive structured, reliable answers instantly.

At its core, the server offers a collection of well‑defined tools that translate natural language requests into concrete map queries. Each tool is built on top of freely available services—OSRM for routing, Overpass for node and relation data, and Nominatum for geocoding—ensuring that the responses are both accurate and up‑to‑date. The API surface includes address resolution, reverse geocoding, distance calculations between points or addresses, and filters for specific amenities such as parking lots, toilets, post offices, cafés, fast‑food venues, and even Irish street names. An additional tool exposes raw Overpass query results, giving advanced users the flexibility to craft custom searches.

For developers integrating AI assistants into their workflows, Púca provides a turnkey solution. The included Docker compose configuration spins up both the MCP server and a Streamlit UI, allowing quick prototyping or demoing. The Python example using demonstrates how to consume the server from code, while instructions for Claude show how to register the MCP endpoint in a desktop client. Because all interactions are performed over SSE, latency remains low and the assistant can stream responses as data arrives.

Real‑world use cases abound: a navigation app could let users ask an AI “Where is the nearest public toilet?” and receive coordinates that can be plotted on a map; an emergency response system could query “Show me all defibrillators within 2 km of this address” to aid dispatch; a travel planner might ask for “Irish street names near the city center” to generate culturally relevant itineraries. The server’s ability to combine multiple tools—e.g., find a café, then calculate the distance from the user’s location—enables sophisticated conversational flows without burdening the assistant with external API management.

Púca’s standout advantage is its open‑source, volunteer‑driven foundation. By leveraging OpenStreetMap data and providing a clean MCP interface, it offers developers a low‑cost, high‑value resource that scales with their needs. Whether you’re building a chatbot, a mobile app, or an internal tool, the Osm‑UI‑Puca MCP server turns raw map data into actionable knowledge that an AI assistant can deliver with ease.