MCP Server

Connect your location data to any AI assistant

The Model Context Protocol (MCP) lets AI assistants call tools directly. The Geolocate Me MCP server gives your AI assistant access to your phone’s GPS — just ask "where am I?" and it calls the tool automatically.

No curl commands. No manual API calls. Just natural language.

Geolocate Me ships a hosted remote MCP server over HTTP. You don’t need to install anything locally.

  https://geolocateme.guleki.com/mcp

Authentication uses OAuth 2.0 with PKCE. On first connect, you’ll be redirected to an authorize page where you paste your Device ID and Bot Token. After that, your AI client stores a signed access token and reconnects automatically.

get_current_location
  Get the most recent GPS coordinates of your tracked device.
  Coordinates are reverse-geocoded to a human-readable address.
  No parameters required.

get_location_history
  Get location history with optional time filters. Returns up to 1000 entries.
  Parameters:
    hours (optional) — last N hours of history
    from (optional) — ISO 8601 start timestamp
    to (optional) — ISO 8601 end timestamp

1. Open Claude.ai → Settings → Connectors
2. Click “Add custom connector”
3. Name: Geolocate Me
4. Server URL: https://geolocateme.guleki.com/mcp
5. Click Connect → you’ll be redirected to the authorize page
6. Paste your Device ID + Bot Token from the app

Claude handles OAuth and token refresh automatically after the first authorization.

Claude Desktop supports remote MCP servers via its Connectors UI:

1. Open Settings → Developer
2. Add a custom connector
3. URL: https://geolocateme.guleki.com/mcp
4. Complete the OAuth flow when prompted

Alternatively, for config-file-based setups, use the URL form:

{
  "mcpServers": {
    "geolocate-me": {
      "type": "http",
      "url": "https://geolocateme.guleki.com/mcp"
    }
  }
}

Add to your project’s .mcp.json or use the CLI:

  claude mcp add --transport http geolocate-me https://geolocateme.guleki.com/mcp

Or manually:

{
  "mcpServers": {
    "geolocate-me": {
      "type": "http",
      "url": "https://geolocateme.guleki.com/mcp"
    }
  }
}

Claude Code will open the OAuth authorize page on first connect.

Any MCP client that supports remote HTTP servers with OAuth can connect using the same URL:

  https://geolocateme.guleki.com/mcp

Config format varies by client, but the pattern is the same — add the server URL and complete the OAuth flow.

Example (Cursor / Windsurf):

{
  "mcpServers": {
    "geolocate-me": {
      "type": "http",
      "url": "https://geolocateme.guleki.com/mcp"
    }
  }
}

ChatGPT: add as a custom connector from Settings → Apps & connectors.

Geolocate Me is also listed on MCP discovery registries:

• smithery.ai — one-click install for supported clients
• mcp.so — searchable catalog

Both will route you through the same OAuth flow as a direct URL install.

Once connected, ask your AI assistant:

• "Where am I right now?"
• "Where did I park?"
• "Show my location history for the last 6 hours"
• "Where was I yesterday between 9am and 5pm?"
• "How long was I at the office today?"
• "Draft a travel log from my pings this week"

Responses include reverse-geocoded street addresses, not just raw coordinates.

In the Geolocate Me app:

1. Open Settings (gear icon on the dashboard)
2. Tap "Copy All Credentials"
3. Paste into the OAuth authorize page when prompted

Your Device ID and Bot Token are issued once at registration. They authorize the MCP server to read your location history — nothing else.

For advanced clients or custom integrations:

• Authorization server metadata
  GET /.well-known/oauth-authorization-server

• Protected resource metadata
  GET /.well-known/oauth-protected-resource

• Dynamic client registration (RFC 7591)
  POST /oauth/register

• Authorization endpoint
  /oauth/authorize

• Token endpoint
  POST /oauth/token

All endpoints are rooted at https://geolocateme.guleki.com and support PKCE (S256).