home / mcp / korea weather mcp server

Korea Weather MCP Server

Provides access to Korea Meteorological Administration weather data via MCP, enabling grid location lookups and short-term forecasts for Korea.

Installation
Add the following to your MCP client configuration file.

Configuration

View docs
{
  "mcpServers": {
    "jikime-py-mcp-ko-weather": {
      "command": "mcp",
      "args": [
        "run",
        "src/server.py"
      ],
      "env": {
        "KO_WEATHER_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

You can access Korea Meteorological Administration weather data through this MCP server, enabling AI agents to fetch precise short-term forecasts for any location in Korea. It serves structured weather data optimized for consumption by language models and other MCP clients, covering coordinates lookup, forecasts, and key weather attributes.

How to use

Use an MCP client to connect to this server and request weather information for a specific location. First determine the exact grid coordinates for your target area, then fetch the ultra-short-term forecast for that location. You will receive a structured weather report that includes temperature, sky condition, precipitation, humidity, wind, and hourly updates.

How to install

Prerequisites: Python 3.12+ and an API key from the Korea Meteorological Administration.

1. Set up the project locally by obtaining the source files and dependencies. The server can be started from the local environment using MCP tooling.

2. Create a virtual environment and install dependencies using the installed MCP tooling.

3. Provide your API credentials in a configuration file.

4. Run the MCP server locally.

Additional configuration and notes

Environment variables you may need to set include KO_WEATHER_API_KEY for accessing the KMA API. You can place credentials in a dedicated .env file that the server reads on startup.

Two common runtime commands are used to operate the server and its inspector during development: the server run command and the MCP Inspector. These are executed via the MCP command-line tool.

Tools and endpoints

The server exposes tools to obtain grid locations and forecasts, plus resources for weather instructions and prompts to guide queries. These tools are designed to deliver clear, structured weather data suitable for processing by language models.

Examples of usage scenarios

- Retrieve the grid coordinates for 서울특별시 서초구 양재1동 to query precise forecasts.

- Request an ultra-short-term forecast for a specific neighborhood and receive hourly updates with temperature, sky condition, and wind information.

Security and maintenance

Protect your API key and rotate credentials as needed. Store credentials in a secure, non-public location and limit access to the server process.

Troubleshooting

If you cannot connect, verify that your API key is valid, the server is running, and the request parameters include a valid city, gu, and dong. Check the server logs for any authentication or data-fetch errors.

API Reference

Get Grid Location and Get Forecast are the primary tools you will use to translate location information into coordinates and then obtain weather data. Each tool returns structured results tailored for MCP clients.

Available tools

Get Grid Location

Retrieves the grid coordinates nx, ny for a specified Korean city, district, and neighborhood, enabling precise location targeting for the KMA API.

Get Forecast

Fetches the ultra-short-term weather forecast for a specific location using the KMA API, including temperature, precipitation, sky condition, humidity, wind direction, and wind speed.

Weather Instructions

Provides detailed usage guidelines and response formats for weather-related queries.

Prompts

Structured prompt templates to guide conversations and ensure clear forecast presentation.