Schwaizer

MCP Server: Geo MCP

An unofficial MCP server for interacting with Swiss building and geographic data via geo.admin.ch.

This is a community project by Schwaizer and is not an official implementation by the Swiss government.

About Schwaizer

SHAPING SWITZERLAND'S AI FUTURE
Empowering Swiss businesses and society through responsible AI adoption.
Founded in 2025, Schwaizer is a non-profit organization dedicated to accelerating the responsible adoption of artificial intelligence across Switzerland.

Schwaizer Geo MCP Server

MCP (Model Context Protocol) server providing access to Swiss building registry and geography from geo.admin.ch, with optional company-location enrichment (via Zefix MCP).

Features

Building identification and search (BFS GWR: ch.bfs.gebaeude_wohnungs_register)
Coordinate conversions (WGS84 ↔ LV95, LV03 ↔ LV95)
Spatial analytics (radius searches, area statistics)
Company-location integration (companies registered at a building, company → buildings)
Unified missing data semantics (see Data Model)

Data Model & Missing Values (Important)

To ensure consistent downstream processing across tools, the server uses the following conventions:

Missing/unknown values are returned as null.
Coordinates are numeric (LV95 easting/northing). When coordinates are not available, fields are null.
Postal code is numeric when available, otherwise null.
Category, floors, dwellings, area, volume are numeric when present; otherwise null.
When upstream sources do not publish an attribute (e.g., some buildings have no ganzwhg/dwellings or gvol*/volume), the server returns null for that field (no estimation by default).

Example normalized building object snippet:

{
  "egid": "2372938",
  "address": "Bahnhofquai 15",
  "municipality": "Zürich",
  "canton": "ZH",
  "postalCode": 8001,
  "city": "Zürich",
  "buildingCategory": 1060,
  "constructionYear": 1894,
  "numberOfFloors": 8,
  "numberOfDwellings": null,
  "area": 1229,
  "volume": null,
  "coordinates": { "easting": 2683276.64, "northing": 1247892.09 }
}

Available Tools

Coordinate Tools

wgs84_to_lv95
Convert latitude/longitude (WGS84, EPSG:4326) → LV95 (EPSG:2056).

Arguments:

{ "latitude": 47.378177, "longitude": 8.540192 }

Returns:

{ "easting": 2683187.XX, "northing": 1248065.XX, "system": "LV95" }

lv95_to_wgs84
Convert LV95 → WGS84.

Arguments:

{ "easting": 2683188.02, "northing": 1248065.98 }

Returns:

{ "latitude": 47.378177, "longitude": 8.540192, "system": "WGS84" }

lv03_to_lv95
Grid‑based precise conversion (LV03 → LV95).

Arguments:

{ "easting": 683187.12, "northing": 248066.14 }

lv95_to_lv03
Grid‑based precise conversion (LV95 → LV03).

Arguments:

{ "easting": 2683188.02, "northing": 1248065.98 }

Building Tools (BFS GWR)

identify_building
Identify buildings at LV95 coordinates.

Arguments:

{
  "easting": 2683260.75,
  "northing": 1247919.125,
  "tolerance": 12,
  "lang": "en"
}

Returns: array of normalized building objects (see Data Model), with missing values = null.

search_building
Search a building by address or location name.

Arguments:

{
  "address": "Bahnhofplatz 1, 8001 Zürich",
  "canton": "ZH",
  "lang": "en"
}

Returns:

{
  "found": 1,
  "buildings": [ { "egid": "2372938", "... normalized fields ..." } ]
}

get_building_details
Get normalized building details by EGID.

Arguments:

{ "egid": "263016183", "lang": "en" }

Returns (example):

{
  "egid": "263016183",
  "address": "Hausmatt 508",
  "municipality": "Schlossrued",
  "canton": "AG",
  "buildingCategory": 1020,
  "constructionYear": 2016,
  "numberOfFloors": 4,
  "numberOfDwellings": 8,
  "area": 269,
  "volume": null,
  "coordinates": { "easting": 2649057.8, "northing": 1237971.7 }
}

Spatial Analytics

find_buildings_in_radius
Find buildings near a WGS84 point with optional filters.

Arguments:

{
  "lat": 47.378177,
  "lon": 8.540192,
  "radius_km": 0.5,
  "filters": {
    "constructionYearMin": 2000,
    "minFloors": 3
  }
}

Returns summary with buildings[], each entry includes address, EGID (if found), coordinates (WGS84+LV95), distance_m, and key attributes.

get_area_building_statistics
Aggregate statistics for a municipality, canton, or bbox.

Arguments:

{
  "bbox": {
    "minLat": 47.376,
    "minLon": 8.538,
    "maxLat": 47.380,
    "maxLon": 8.543
  }
}

Returns:

totalBuildings
constructionYearHistogram (decade bins)
buildingCategoryBreakdown (counts + percentages)
avgFloors
oldestBuilding, newestBuilding
residentialShare (computed based on category ranges; may include mixed use per BFS categories)

Company‑Location Tools (optional)

These tools enrich building data with company information. They can operate standalone using the public UID Webservice; for best results, run Schwaizer Zefix MCP alongside and configure its URL in your environment if applicable.

get_building_companies
List companies registered at a building address.

Arguments:

{ "egid": 2372938 }

Returns:

{
  "egid": 2372938,
  "address": { "street": "Bahnhofquai", "houseNumber": "15", "postalCode": 8001, "city": "Zürich" },
  "totalCompanies": 5,
  "companies": [
    { "uid": "CHE-106.018.335", "name": "Nextlane Switzerland AG", "legalForm": "Corporation (AG)", "status": "ACTIVE" }
    // ...
  ]
}

find_company_buildings
Resolve all buildings (HQ + branches) for a company UID.

Arguments:

{ "companyUid": "CHE-106.018.335" }

Returns:

{
  "company": { "uid": "CHE-106.018.335", "name": "...", "legalForm": "...", "status": "..." },
  "locations": [
    {
      "locationType": "soap",
      "address": "Bahnhofquai 11, 8001 Zürich",
      "egid": "2372938",
      "coordinates": { "wgs84": { "lat": 47.37660, "lon": 8.54133 }, "lv95": { "easting": 2683276.64, "northing": 1247892.09 } },
      "buildingDetails": { "constructionYear": 1894, "buildingCategory": 1060, "floors": 8 }
    }
  ]
}

Note: Debug details were intentionally removed from the response for cleaner outputs.

enrich_building_with_companies
Return building info plus detailed company data.

Arguments:

{ "egid": 2372938 }

analyze_building_occupancy
Heuristic classification of occupancy type using category + presence of companies.

Arguments:

{ "egid": 2372938 }

Returns:

{
  "egid": 2372938,
  "address": "Bahnhofquai 15, 8001 Zürich",
  "buildingCharacteristics": { "category": 1060, "categoryType": "residential", "isMixedUse": false, "floors": 8, "dwellings": null, "constructionYear": 1894 },
  "commercialOccupancy": { "totalCompanies": 5, "hasCommercialActivity": true, "companiesPerFloor": "0.63" },
  "occupancyType": "residential_with_commercial"
}

Installation

npm install

Quickstart with Schwaizer Zefix MCP (Option 1 — Sibling directories)

Schwaizer Geo MCP can enrich building data with company information via the Schwaizer Zefix MCP. The simplest setup is to clone both repositories side‑by‑side and use the local dependency already defined in package.json.

Directory layout:

parent/
  schwaizer-zefix-mcp/
  schwaizer-geo-mcp/

Why this works:

Schwaizer Geo MCP’s package.json references the Zefix MCP locally:
```
"schwaizer-zefix-mcp": "file:../schwaizer-zefix-mcp"
```
Placing both repos as siblings allows npm install in Schwaizer Geo MCP to resolve the Zefix MCP locally without extra configuration.

Steps:

Clone both repositories

# from your parent directory
git clone https://github.com/schwaizer/schwaizer-zefix-mcp.git
git clone https://github.com/schwaizer/schwaizer-geo-mcp.git

Install & configure Zefix MCP first

cd schwaizer-zefix-mcp
cp .env.example .env
# Optional: only needed for Zefix REST enrichment (UID Public Services need no auth)
# ZEFIX_USERNAME=your_username
# ZEFIX_PASSWORD=your_password
# UID_PUBLIC_URL=https://www.uid-wse.admin.ch/V5.0/PublicServices.svc?wsdl
npm install

Install Schwaizer Geo MCP

cd ../schwaizer-geo-mcp
cp .env.example .env
# Optional: adjust
# GEOADMIN_BASE_URL=https://api3.geo.admin.ch
# CACHE_TTL=300000
# LOG_LEVEL=info
npm install

Run both servers (separate terminals)

# Terminal 1
cd schwaizer-zefix-mcp
npm start

# Terminal 2
cd schwaizer-geo-mcp
npm start

Add one MCP server entry for Zefix
Add a second MCP server entry for Schwaizer Geo MCP
You’ll see both tool catalogs available. Schwaizer Geo MCP can enrich buildings with company info, and you can also call Zefix tools directly.

Validation

From schwaizer-geo-mcp, npm ls schwaizer-zefix-mcp should resolve to the sibling schwaizer-zefix-mcp folder.
If it doesn’t resolve (wrong paths/structure), use the same parent folder layout or consider npm link/workspaces.

Notes

Zefix credentials are optional: UID Public Services require no auth; Zefix REST enrichments (e.g., purpose/registration date) require credentials.
All missing values are returned as null (never "N/A"), consistent across both servers.

Configuration

Copy .env.example to .env and adjust:

# Optional: Custom base URL for geo.admin.ch API
GEOADMIN_BASE_URL=https://api3.geo.admin.ch

# Optional: Cache TTL in milliseconds (default: 5 minutes)
CACHE_TTL=300000

# Log level (trace|debug|info|warn|error|fatal), default: info
LOG_LEVEL=info

If you intend to integrate with Schwaizer Zefix MCP for richer company data, run it separately and configure a URL/environment according to that server’s README. This server primarily queries UID Public Services directly for addresses and enriches via Zefix REST when available.

Usage

Running the Server

npm start

Example Tool Calls

Identify building:

{
  "name": "identify_building",
  "arguments": {
    "easting": 2683260.75,
    "northing": 1247919.125,
    "tolerance": 12,
    "lang": "en"
  }
}

Search building:

{
  "name": "search_building",
  "arguments": {
    "address": "Bahnhofplatz 1, 8001 Zürich",
    "canton": "ZH",
    "lang": "en"
  }
}

Find buildings in radius:

{
  "name": "find_buildings_in_radius",
  "arguments": {
    "lat": 47.378177,
    "lon": 8.540192,
    "radius_km": 0.5,
    "filters": { "constructionYearMin": 2000, "minFloors": 3 }
  }
}

Get area stats:

{
  "name": "get_area_building_statistics",
  "arguments": {
    "bbox": { "minLat": 47.376, "minLon": 8.538, "maxLat": 47.380, "maxLon": 8.543 }
  }
}

Get building companies:

{
  "name": "get_building_companies",
  "arguments": { "egid": 2372938 }
}

API Documentation & Sources

geo.admin.ch services:
- API overview: https://api3.geo.admin.ch/services/sdiservices.html
- Search API: https://api3.geo.admin.ch/rest/services/api/SearchServer
- MapServer Identify: https://api3.geo.admin.ch/rest/services/api/MapServer/identify
- Height services: https://api3.geo.admin.ch/rest/services/height

Caching & Rate Limits

Default in-memory cache TTL is 5 minutes (CACHE_TTL).
geo.admin.ch endpoints are public and rate-limited; caching reduces repeated requests.
Cache resets on server restart.

Development

Project Structure

  schwaizer-geo-mcp/
├── src/
│   ├── index.js                  # MCP server entry point
│   ├── config.js                 # Configuration loader
│   ├── api/
│   │   ├── geoadmin-client.js    # geo.admin.ch client
│   │   └── zefix-clients.js      # UID/Zefix clients (integration helpers)
│   ├── tools/
│   │   ├── building-tools.js     # identify/search/details
│   │   ├── coordinate-tools.js   # conversions
│   │   ├── spatial-analytics.js  # radius & area stats
│   │   └── company-location.js   # company & occupancy tools
│   └── utils/
│       ├── logger.js             # logging
│       └── cache.js              # caching
├── tests/
│   ├── setup.js
│   ├── unit/
│   └── integration/
├── docs/
│   └── CHANGELOG.md
├── .env.example
├── eslint.config.js
├── vitest.config.js
├── package.json
└── README.md

Running Tests

npm test

Linting

npm run lint

License

MIT

Support

geo.admin.ch: https://api3.geo.admin.ch/
This MCP Server: Open an issue on GitHub

ishumilin/schwaizer-geo-mcp

Schwaizer

MCP Server: Geo MCP

About Schwaizer

Schwaizer Geo MCP Server

Features

Data Model & Missing Values (Important)

Available Tools

Coordinate Tools

Building Tools (BFS GWR)

Spatial Analytics

Company‑Location Tools (optional)

Installation

Quickstart with Schwaizer Zefix MCP (Option 1 — Sibling directories)

Configuration

Usage

Running the Server

Example Tool Calls

API Documentation & Sources

Caching & Rate Limits

Development

Project Structure

Running Tests

Linting

License

Support

On this page

Languages

Contributors