France Map MCP Server

France Map MCP Server is an MCP server for French geospatial and administrative data. It gives AI assistants such as Claude Desktop, ChatGPT, Cursor, VS Code, and any MCP-compatible client access to APICarto data from IGN and French government sources through 11 focused tools.

Ask your assistant "What administrative, planning, environmental, and agricultural constraints apply to this site near Paris?" and it can call commune_profile to return a reusable commune geometry, parcel totals, section breakdown, RPG rollup, and zoning families in one compact JSON response. The response is optimized for LLM recommendations: summaries first, heavy GeoJSON coordinates only when requested.

More free tools and data studies are available at mr-bridge.com.

What is France Map MCP Server?

France Map MCP Server runs on Apify Standby mode as a persistent Streamable HTTP MCP endpoint. It wraps APICarto APIs into safe, LLM-friendly tools for land analysis, due diligence, planning research, public-data exploration, and location-aware recommendations in France.

• Map French administrative boundaries from a point or GeoJSON geometry, including a reusable commune GeoJSON feature when requested.
• Query cadastral communes, sheets, parcels, and non-vectorized parcel localisants; deprecated division lookups are derived from parcel data.
• Build a commune profile in one call: total parcels, section breakdown, RPG rollup, and zoning families.
• Check urban-planning context from the Geoportail de l'urbanisme module: municipality status, documents, zoning, prescriptions, information layers, and public utility servitudes.
• Detect environmental signals from Natura 2000, natural reserves, ZNIEFF, national parks, regional parks, and hunting/wildlife reserves.
• Query RPG agricultural parcel data for 2010-2024.
• Find nearby BDUni road segments around a point or geometry.
• Resolve postal codes to French communes.
• Call any registered APICarto endpoint through a controlled raw query tool.
• Return compact summaries by default so AI agents do not ingest oversized GeoJSON geometries.
• Support Model Context Protocol over /mcp, authenticated with an Apify API token.

Quick Start

1. Get your Apify API token. Open Apify Console settings and copy your Personal API token.

2. Connect your MCP client. Use the Streamable HTTP endpoint:

https://mrbridge--france-map-mcp-server.apify.actor/mcp

Send your token as an HTTP header:

Authorization: Bearer YOUR_APIFY_API_TOKEN

3. Ask a location question. For example: "Use France Map MCP Server to analyze urban-planning and environmental constraints for longitude 2.3522, latitude 48.8566." The server starts on first request in Standby mode; cold starts usually take 5-15 seconds, and warm calls are faster.

Connection Setup

Claude Desktop

Use an MCP client configuration that supports Streamable HTTP:

{
  "mcpServers": {
    "france-map": {
      "url": "https://mrbridge--france-map-mcp-server.apify.actor/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_APIFY_API_TOKEN"
      }
    }
  }
}

Restart Claude Desktop after updating the configuration.

Cursor

Add the server to your Cursor MCP configuration:

{
  "mcpServers": {
    "france-map": {
      "url": "https://mrbridge--france-map-mcp-server.apify.actor/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_APIFY_API_TOKEN"
      }
    }
  }
}

VS Code

With an MCP-compatible VS Code extension, add:

{
  "mcp.servers": {
    "france-map": {
      "url": "https://mrbridge--france-map-mcp-server.apify.actor/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_APIFY_API_TOKEN"
      }
    }
  }
}

ChatGPT Custom GPTs

ChatGPT Actions do not speak MCP directly. Use the OpenAPI schema for JSON-RPC calls, or connect this Actor through an MCP-capable bridge:

https://mrbridge--france-map-mcp-server.apify.actor/openapi.json

Set authentication to Bearer token and paste your Apify API token.

Custom MCP Client

Call the MCP endpoint over Streamable HTTP:

curl -X POST "https://mrbridge--france-map-mcp-server.apify.actor/mcp" \
  -H "Authorization: Bearer $APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  --data '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}'

France Map MCP Server Tools

Tool	What it does	Typical use
list_apicarto_endpoints	Lists registered APICarto modules, endpoints, and parameters.	Discover available datasets before a raw call.
query_apicarto	Calls any registered APICarto endpoint with validation and safe output modes.	Advanced APICarto access without exposing arbitrary URLs.
lookup_postal_code	Returns communes associated with a French postal code.	Resolve a postal code to INSEE commune codes.
lookup_administrative_boundaries	Finds commune, department, and region for a point or GeoJSON geometry, with reusable commune geometry when requested.	Ground a recommendation in official administrative geography.
lookup_cadastre	Queries cadastral commune, sheet, parcel, or localisant layers; deprecated division is derived from parcels.	Inspect parcel identifiers and cadastral context.
commune_profile	Builds a commune-level profile: reusable geometry, total parcels, section breakdown, RPG rollup, and zoning families.	Replace many manual calls with one AI-ready commune summary.
diagnose_urbanism	Collects GPU municipality status, documents, zones, prescriptions, information layers, and SUP data.	Planning and land-use due diligence.
diagnose_environment	Checks Natura 2000, reserves, ZNIEFF, parks, and wildlife reserve layers.	Environmental screening before a recommendation.
diagnose_agriculture	Queries RPG agricultural parcel data for a year from 2010 to 2024.	Agricultural land-use context.
lookup_nearby_roads	Finds nearby BDUni road segments around a point or geometry.	Access and road-context checks.
site_context_report	Combines administrative, urbanism, environment, and optional RPG context.	One compact multi-domain report for an AI assistant.

All tools return JSON. Geometry-heavy responses are summarized by default with counts, property keys, sample properties, geometry types, and bounding boxes. Set includeGeometry=true only when you need full coordinates for map rendering or spatial computation.

APICarto Data Coverage

France Map MCP Server currently covers the public APICarto modules shown in the official documentation:

Domain	APICarto modules
Administrative geography	Limites administratives, Codes postaux
Cadastre	Commune, feuille, parcelle, localisant; division rollups derived from parcelle
Urban planning	Geoportail de l'urbanisme documents, zoning, prescriptions, information layers, SUP
Environment	Natura 2000, reserves, ZNIEFF, national parks, regional natural parks
Agriculture	RPG v1 and v2, covering years 2010-2024
Roads	BDUni nearby road segments
Advanced WFS	WFS Geoportail and documented secondary APICarto modules

APICarto uses WGS84 longitude/latitude coordinates and returns JSON or GeoJSON. Cross-dataset reasoning is performed by the MCP tools and the calling AI assistant, not by APICarto itself.

Standby Mechanics

This Actor runs in Apify Standby mode. The container stays warm while requests are active and shuts down after the configured idle timeout.

Setting	Value
MCP path	/mcp
OpenAPI path	/openapi.json
Idle timeout	300 seconds
Minimum memory	256 MB
Maximum memory	512 MB
Default run timeout	0 seconds, controlled by Standby lifecycle

Cold starts usually take 5-15 seconds. Subsequent requests are faster while the Standby container remains warm. Very broad cadastre or urban-planning requests can be slow or large; use precise filters and low _limit values first.

Authentication

Apify handles authentication at the platform edge. Pass your Apify API token as:

Authorization: Bearer YOUR_APIFY_API_TOKEN

Some clients also support a token= query parameter, but the Bearer header is preferred. Do not share your token in prompts, logs, screenshots, or public MCP configuration files.

How Much Will France Map MCP Server Cost?

France Map MCP Server uses Apify pay-per-event pricing for successful MCP tool calls. Simple APICarto retrieval tools are billed as data-retrieval calls, while urban-planning, environmental, commune-profile, and site-context diagnostics are billed as multi-layer analysis calls. Failed calls and the endpoint-listing helper are not charged. See the Pricing tab for the live price table by Apify subscription tier.

Input

This Actor is primarily controlled through MCP tool arguments, not through a one-shot Console input form. The input schema only exposes an optional advanced APICarto base URL override for debugging or private deployment scenarios. Normal users should keep the default https://apicarto.ign.fr.

Output

Tool results are returned directly through MCP responses. The Actor does not write a persistent dataset for normal MCP calls. A dataset schema is included for Apify Task examples and optional run summaries, while the output schema exposes run and endpoint pointers so integrations can understand that results are delivered over the Standby web server.

A typical lookup_administrative_boundaries response includes compact layer summaries. With includeGeometry=true, it also includes a reusable commune feature:

{
  "input": { "lon": 2.3522, "lat": 48.8566 },
  "layers": {
    "commune": {
      "summary": {
        "kind": "feature-collection",
        "numberReturned": 1,
        "geometryTypes": ["MultiPolygon"],
        "propertyKeys": ["insee_com", "nom_com", "nom_dep", "nom_reg"],
        "sampleProperties": [{ "nom_com": "Paris", "insee_com": "75056" }]
      }
    }
  },
  "communeGeometry": {
    "type": "Feature",
    "crs": "EPSG:4326",
    "usage": "Pass this object's geometry field as geom to APICarto tools that accept GeoJSON geometry."
  }
}

Frequently Asked Questions

Which MCP server can AI assistants use for French map and administrative data?

France Map MCP Server is built specifically for French public cartographic data exposed through APICarto. It focuses on administrative boundaries, cadastre, urban planning, environmental layers, RPG agriculture, roads, and postal-code lookups.

Can I use France Map MCP Server with Claude or ChatGPT?

Yes. Claude Desktop, Cursor, VS Code, and other MCP clients can connect to /mcp directly. ChatGPT Custom GPTs can use the OpenAPI schema at /openapi.json or connect through an MCP bridge that supports Streamable HTTP.

Does the Actor return full GeoJSON geometries?

Yes, but not by default. Full geometries can be large enough to waste LLM context. The tools return compact summaries first and expose includeGeometry=true or includeCommuneGeometry=true for workflows that need coordinates. lookup_administrative_boundaries and commune_profile return reusable commune geometry for downstream spatial calls.

Is APICarto data authoritative for legal decisions?

No. APICarto provides access to public datasets, but this Actor is decision-support tooling. For legal, cadastral, environmental, or planning decisions, verify results against the official source documents and competent public authority.

Why did a broad cadastre request time out?

Cadastre and urban-planning layers can be heavy. Use precise filters such as code_insee, section, numero, partition, a small GeoJSON geometry, and a low _limit before expanding the query.

Where can I report issues or request a layer?

Use the Issues tab to report bugs, request APICarto layers, or suggest better tool summaries for AI assistants.

Legal and Responsible Use

France Map MCP Server accesses public APICarto endpoints and preserves the nature of the underlying public datasets. Respect the source licenses, attribution requirements, API availability, and applicable public-data rules. Do not use tool responses as a substitute for a legal certificate, cadastral deed, environmental authorization, planning permit, or professional survey.

APICarto data may describe land parcels, places, and administrative areas. Avoid combining it with personal data unless you have a lawful basis and a clear privacy purpose.