• RAW DATA
  • JSON
  • SUMMARY
  • HUMAN

How to extract custom geographic data with your AI

Connect the Overpass MCP to your agent to turn natural language requests into precise spatial queries. You get raw map data without writing a single line of code.

Ask AI about this page

Short answer

How do I pull specific map data using my AI?

Your AI uses the custom_query tool to translate your descriptions into Overpass QL. It then pulls the requested nodes, ways, or relations from OpenStreetMap. You'll receive the specific spatial attributes you asked for.

Where every query lands.

Data outcomes

The AI processes your request and delivers the results in these formats.

RAW DATA

Structured spatial layers

The custom_query tool fetches specific OSM elements. You get the raw tags and coordinates for the objects you requested.

JSON

Machine-readable files

Your agent can format the retrieved map data into clean JSON. This makes it easy to move the data into other tools or scripts.

SUMMARY

Human-readable counts

Instead of raw coordinates, your AI can summarize what it found. It tells you how many cafes or parks exist in a specific area.

HUMAN

Manual query refinement

If the query is too broad or the syntax is tricky, you step in. You'll need to refine the search parameters to get the exact layer you want.

The workflow

What your AI does when a query arrives.

The AI handles the heavy lifting of translating intent into map language.

  1. Translate intent

    Your agent interprets your plain English request and converts it into a valid Overpass QL string.

    custom_query

  2. Execute query

    The AI sends the formatted query to the OpenStreetMap engine to find the matching geographic features.

    custom_query

  3. Parse results

    The agent receives the XML or JSON response from the map server and organizes the data.

    custom_query

  4. Format output

    The AI presents the final data as a list, a table, or a structured file depending on your needs.

    custom_query

Try it

Copy these to start.

Use these prompts to trigger your first spatial query.

Starting points

Swap out the bracketed locations or object types for your actual needs.

Overpass API Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_DGkmVxEbxmLxMu1OyGbUJIrzjQi3rn2xOK8B2fYg/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — Overpass API capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "overpass-openstreetmap-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_DGkmVxEbxmLxMu1OyGbUJIrzjQi3rn2xOK8B2fYg/mcp"
    }
  }
}
Copy into chat04

  • Find all the restaurants in downtown Chicago and give me their names and addresses.

  • Run a query to find all the bicycle paths within 5km of Central Park.

  • List all the hospitals in London that have a pharmacy nearby.

  • Get the coordinates for every park in Berlin.

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect Overpass once, then ask.

One click connects the MCP to your client. Your credentials stay encrypted, and you can start querying immediately.

Connect Overpass to your AI

FAQ

How the Overpass MCP behaves

  • 01

    Can the AI edit OpenStreetMap data?

    No. This MCP is read-only. It can only retrieve data that already exists in OpenStreetMap.

  • 02

    Do I need to know Overpass QL to use this?

    No. Your AI handles the translation from your natural language into the required query language.

  • 03

    Can the AI download large map files?

    No. The AI is limited by the tool's ability to process specific queries. It is meant for targeted data extraction, not full map dumps.

  • 04

    Will the AI save my queries?

    The AI client you use manages your chat history, but the MCP itself does not store your previous queries or results.

  • 05

    Can I use this to find private property boundaries?

    No. The AI can only access data that has been publicly contributed to OpenStreetMap.

  • More questions about Overpass? See everything the Overpass Connector can do

Connect Overpass API to Claude, Cursor, ChatGPT & more