OneMap MCP Server v2

OneMap MCP Server v2

A Python-based MCP (Model Context Protocol) server that provides comprehensive access to Singapore's OneMap APIs. Built with FastMCP for easy integration with AI assistants and Microsoft AI Foundry.

Features

This server exposes 35+ tools across 10 API categories:

• Search - Address and location search
• Reverse Geocode - Convert coordinates to addresses (WGS84 and SVY21)
• Routing - Public transport, driving, walking, cycling, barrier-free routes
• Coordinate Converters - EPSG 4326 (WGS84), EPSG 3414 (SVY21), EPSG 3857
• Themes - Access 100+ thematic layers for locations, amenities, boundaries
• Planning Area - Singapore's 55 planning area information
• Population Query - Demographics and statistics by planning area
• Nearby Transport - Find nearby MRT/LRT stations and bus stops
• Static Map - Generate static map images with optional overlays

Prerequisites

• Python 3.11+
• OneMap Account (register at OneMap API)

Installation

Local Development

1. Clone the repository:

cd onemap-mcp

2. Create a virtual environment:

python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Create .env file with your credentials:

cp .env.example .env
# Edit .env with your OneMap credentials

5. Run the server:

python server.py

Environment Variables

Create a .env file with:

[email protected]
ONEMAP_EMAIL_PASSWORD=your_password

Docker Deployment

1. Build the Docker image:

docker build -t onemap-mcp .

2. Run the container:

docker run -d \
  -e ONEMAP_EMAIL="[email protected]" \
  -e ONEMAP_EMAIL_PASSWORD="your_password" \
  --name onemap-mcp \
  onemap-mcp

Available Tools

Search & Geocoding

Routing

Coordinate Conversion

Themes

Planning Areas

Population Data

Transport

Static Maps

Usage Examples

Search for a location

# Search for Marina Bay Sands
result = await search(search_value="Marina Bay Sands")

Get route directions

# Driving route from Changi to Orchard
result = await route_walk_drive_cycle(
    start_lat=1.3644,
    start_lon=103.9915,
    end_lat=1.3048,
    end_lon=103.8318,
    route_type="drive"
)

Find nearby MRT stations

result = await get_nearby_mrt_stations(
    latitude=1.3521,
    longitude=103.8198,
    radius_in_meters=1000
)

Project Structure

onemap-mcp/
├── server.py          # FastMCP server with all tools
├── mcp.json           # MCP manifest
├── tools.json         # Tool definitions for AI Foundry
├── onemap/
│   ├── __init__.py
│   └── utils.py       # HTTP client and utility functions
├── .env               # Your credentials (not in git)
├── .env.example       # Template for credentials
├── Dockerfile
├── requirements.txt
└── README.md

Deployment to Azure

Azure Container Apps

# Build and push to Azure Container Registry
az acr build --registry <registry-name> --image onemap-mcp:latest .

# Deploy to Container Apps
az containerapp create \
  --name onemap-mcp \
  --resource-group <resource-group> \
  --image <registry-name>.azurecr.io/onemap-mcp:latest \
  --env-vars ONEMAP_EMAIL=<email> ONEMAP_EMAIL_PASSWORD=<password>

Microsoft AI Foundry Integration

1. Deploy the server to a publicly accessible endpoint
2. Use the tools.json file to configure tool definitions
3. Configure ONEMAP_EMAIL and ONEMAP_EMAIL_PASSWORD environment variables

License

MIT

Resources

• OneMap API Documentation
• Model Context Protocol
• Microsoft AI Foundry