Singapore Open Data MCP Server

An MCP (Model Context Protocol) server that provides AI agents with programmatic access to Singapore's government open data from data.gov.sg. It enables querying real-time and historical datasets including weather, transport, property, and air quality.

Key Features

The server exposes several tools for AI agents to interact with Singapore's open data:

• sg.weather_forecast: Fetches the 2-hour weather forecast for all areas in Singapore. No parameters required.
• sg.carpark_availability: Gets real-time parking lot availability. Accepts an optional limit parameter (default: 20).
• sg.taxi_availability: Retrieves live taxi locations and total count. No parameters required.
• sg.psi: Returns current air quality (PSI, PM2.5, PM10) readings by region. No parameters required.
• sg.query_dataset: A generic tool to query any dataset from the platform using its datasetId. Supports optional limit and full-text search (q) parameters.
• sg.get_popular_datasets: Lists commonly used datasets (e.g., HDB Resale Prices, COE Results) with their resource IDs. No parameters required.

How to Use

Connect to the server from any MCP-compatible client (like Claude Desktop) using Streamable HTTP transport.

Endpoint: https://YOUR_USERNAME--sg-open-data-mcp.apify.actor/mcp

Authentication: Requires an Authorization: Bearer YOUR_APIFY_API_TOKEN header.

Example Configuration for Claude Desktop

Add the following to your claude_desktop_config.json:

{
  "mcpServers": {
    "sg-open-data": {
      "url": "https://YOUR_USERNAME--sg-open-data-mcp.apify.actor/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_APIFY_API_TOKEN"
      }
    }
  }
}

Input/Output Examples

Tool: sg.weather_forecast
Input: {}
Output (example):

{
  "updated": "2024-01-15T14:00:00+08:00",
  "forecasts": [
    { "area": "Ang Mo Kio", "forecast": "Partly Cloudy" },
    { "area": "Bedok", "forecast": "Light Showers" }
  ]
}

Tool: sg.carpark_availability
Input: { "limit": 5 }
Output (example):

{
  "timestamp": "2024-01-15T14:30:00+08:00",
  "carparks": [
    {
      "carpark_number": "ACB",
      "info": [{ "lot_type": "C", "total_lots": "105", "available": "23" }]
    }
  ]
}

Tool: sg.query_dataset
Input: { "datasetId": "d_8b84c4ee58e3cfc0ece0d773c8ca6abc", "limit": 10 }
Output: Returns the specific dataset's records and field metadata.

Once connected, an AI agent can use natural language to trigger these tools—for example, asking "What's the weather forecast for Tampines?" will invoke the sg.weather_forecast tool and parse the response.