au-weather-mcp

Australian Weather MCP Server

An MCP (Model Context Protocol) server that provides access to Australian weather data from the Bureau of Meteorology (BOM). This server allows AI assistants and other MCP clients to search for locations and retrieve weather forecasts and current observations for locations across Australia.

Features

This MCP server provides the following tools:

• search_location: Search for Australian locations by suburb name or postcode
• get_daily_forecast: Get multi-day weather forecast for a location
• get_three_hourly_forecast: Get detailed 3-hourly forecast data
• get_observations: Get current weather observations
• get_weather_for_location: Get comprehensive weather info (convenience tool that combines search + forecast + observations)

Installation

Using pip

pip install -e /path/to/au-weather-mcp

Using uv (recommended)

uv pip install -e /path/to/au-weather-mcp

Using Docker Desktop

If you prefer to use Docker, you can build and run the server in a container:

# Clone the repository
git clone git@github.com:craigles75/au-weather-mcp.git
cd au-weather-mcp

# Build the Docker image
docker build -t au-weather-mcp .

# Or use docker compose
docker compose build

Configuration

Claude Desktop

Add this to your Claude Desktop configuration file:

MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json

Windows: %APPDATA%/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "au-weather": {
      "command": "python",
      "args": ["/path/to/au-weather-mcp/server.py"]
    }
  }
}

Or if you're using uv:

{
  "mcpServers": {
    "au-weather": {
      "command": "uv",
      "args": ["--directory", "/path/to/au-weather-mcp", "run", "server.py"]
    }
  }
}

Or if you're using Docker:

{
  "mcpServers": {
    "au-weather": {
      "command": "docker",
      "args": ["run", "-i", "--rm", "au-weather-mcp"]
    }
  }
}

Note: When using Docker, make sure you've built the image first with docker build -t au-weather-mcp . or docker compose build.

Other MCP Clients

For other MCP clients, configure them to run:

Python: ``bash python /path/to/au-weather-mcp/server.py ``

Docker: ``bash docker run -i --rm au-weather-mcp ``

The server communicates over stdio (standard input/output).

Usage Examples

Once configured, you can use the tools through your MCP client (like Claude Desktop):

Search for a location

Use the search_location tool to find "Melbourne"

This returns a list of locations matching "Melbourne" with their geohash identifiers.

Get weather for a location

Use the get_weather_for_location tool for "Sydney"

This searches for Sydney and returns both the daily forecast and current observations.

Get detailed forecast

First search for "Parkdale 3195", then use the geohash to get the daily forecast

This gives you more control over which specific location you want when there are multiple matches.

Tool Reference

search_location

Search for Australian locations by suburb name or postcode.

Parameters:

• query (string, required): Suburb name (e.g., "Parkdale", "Sydney") or postcode (e.g., "3195", "2000")

Returns: List of matching locations with geohash identifiers

get_daily_forecast

Get daily weather forecast for a location.

Parameters:

• geohash (string, required): Location geohash identifier (get from search_location)

Returns: Multi-day forecast with temperatures, conditions, rain predictions

get_three_hourly_forecast

Get detailed three-hourly weather forecast.

Parameters:

• geohash (string, required): Location geohash identifier (get from search_location)

Returns: Detailed 3-hourly forecast data

get_observations

Get current weather observations.

Parameters:

• geohash (string, required): Location geohash identifier (get from search_location)

Returns: Current weather conditions including temperature, humidity, wind, etc.

get_weather_for_location

Get comprehensive weather information (convenience tool).

Parameters:

• location (string, required): Suburb name or postcode

Returns: Location info, daily forecast, and current observations

Development

Running Tests

# Install dev dependencies
pip install -e ".[dev]"

# Run tests
pytest

Testing the Server Manually

You can test the server by running it directly:

Python: ``bash python server.py ``

Docker: ``bash docker run -i --rm au-weather-mcp ``

The server will start and communicate over stdio. You can send MCP protocol messages to test the tools.

Building and Running with Docker

# Build the image
docker build -t au-weather-mcp .

# Run the server
docker run -i --rm au-weather-mcp

# Or using docker compose
docker compose build
docker compose run --rm au-weather-mcp

Data Source

This server retrieves data from the Bureau of Meteorology's public API at https://api.weather.bom.gov.au/v1.

License

MIT

Contributing

Contributions are welcome! Please feel free to submit issues or pull requests.