hk-climate

This MCP server exposes climate and weather data from the Hong Kong Observatory via a FastMCP-like interface. It provides tools to fetch current weather, a 9-day forecast, local forecasts, weather warnings, special tips, visibility and lightning data, moon times, tide information, and weather/radiation reports. Clients can query these data sets through the standardized MCP endpoints exposed by the server and receive structured dictionaries containing data such as temperature, humidity, forecast details, warning messages, and update times. The server is designed to be run locally or in a containerized environment and can be connected to by MCP clients via standard input/output or SSE, depending on the chosen transport mode.

Prerequisites:

• Python 3.8+ installed on your system
• Internet access to install dependencies

Install and run locally:

1. Clone the repository: git clone https://github.com/hkopenai/hk-climate-mcp-server.git cd hk-climate-mcp-server

2. Install Python dependencies: python -m pip install --upgrade pip pip install -r requirements.txt

3. Run the server (stdio mode): python server.py

4. (Optional) Run in SSE mode on a specific host/port: TRANSPORT_MODE=sse HOST=0.0.0.0 PORT=8080 python server.py

5. Connect via Cline or an MCP client using the provided configuration (see Cline Integration in the README).

• The data source for this MCP server is the Hong Kong Observatory (HKO).
• Available APIs include current weather, 9-day forecasts, local forecasts, weather warnings, special weather tips, visibility data, lightning data, moon times, hourly tides, high/low tides, and weather/radiation reports.
• Environment variables control transport mode and binding: TRANSPORT_MODE (stdio or sse), HOST, and PORT. If TRANSPORT_MODE is not set, the server defaults to stdio mode.
• For Cline integration, the example shows how to connect using stdio by running the module with -m hkopenai.hk_climate_mcp_server; ensure the module path matches your installation.
• If you modify config or add additional endpoints, ensure your MCP client and documentation stay in sync with the exposed methods and their language defaults (en/tc/sc, etc.).