MCP Server AccuWeather Hourly and Daily Forecasts

Overview

This MCP (Model Context Protocol) server exposes AccuWeather hourly and daily forecasts as MCP-compatible tools. It bridges AccuWeather’s REST API and model-driven workflows so language models or other MCP-aware agents can request weather data by location and receive consistent, structured forecasts. The server handles API key management, location lookup, result formatting, and rate-limit smoothing so models and integrations get concise, reliable weather context.

For developers, the server is useful when you need location-based weather data inside a model-driven application, a chatbot, scheduling assistant, or automation pipeline. By providing hourly and daily forecasts as MCP tools, you can invoke precise, up-to-date weather conditions without embedding API keys in multiple places or writing custom parsing logic for forecast responses.

Features

• Exposes AccuWeather hourly and daily forecasts as MCP tools
• Location-based lookup: supports lat/lon or AccuWeather location keys
• Structured JSON output formatted for model consumption
• Simple configuration using environment variables or a .env file
• Docker-ready and runnable locally for development
• Example tool manifest for easy integration with MCP-capable models
• Rate-limit and error handling to surface consistent messages to clients

Installation / Configuration

Clone the repository, install dependencies, configure your AccuWeather API key, and start the server.

1. Clone the repo

git clone https://github.com/TimLukaHorstmann/mcp-weather.git
cd mcp-weather

2a. Run locally (Node.js example)

# install
npm install

# set environment variables (or add them to a .env file)
export ACCUWEATHER_API_KEY="your_accuweather_api_key"
export PORT=3000

# start the server
npm start

2b. Run with Docker

# build (if a Dockerfile is included)
docker build -t mcp-weather .

# run with environment variables
docker run -e ACCUWEATHER_API_KEY="your_api_key" -e PORT=3000 -p 3000:3000 mcp-weather

3. Example .env

ACCUWEATHER_API_KEY=your_accuweather_api_key
PORT=3000
DEFAULT_UNITS=metric # or imperial
DEFAULT_LANG=en-us

Configuration options (typical environment variables)

• ACCUWEATHER_API_KEY — required API key for AccuWeather
• PORT — server listen port (default 3000)
• DEFAULT_UNITS — metric or imperial
• DEFAULT_LANG — language code for forecasts (e.g., en-us)
• RATE_LIMIT_WINDOW / RATE_LIMIT_MAX — optional rate-limit settings

Available Tools / Resources

The server exposes MCP-style tools (HTTP endpoints) for forecasts. Typical endpoints and their purpose:

Request payloads are JSON. Example input schema:

{
  "lat": 40.7128,
  "lon": -74.0060,
  "locationKey": null,
  "hours": 24,
  "days": 7,
  "units": "metric",
  "lang": "en-us"
}

Example curl calls:

# Hourly forecast by lat/lon
curl -X POST http://localhost:3000/tools/accuweather/hourly \
  -H "Content-Type: application/json" \
  -d '{"lat":40.7128,"lon":-74.0060,"hours":12}'

# Daily forecast by location key
curl -X POST http://localhost:3000/tools/accuweather/daily \
  -H "Content-Type: application/json" \
  -d '{"locationKey":"349727"}'

Example MCP tool manifest (JSON) to register with an MCP-capable model:

[
  {
    "name": "accuweather_hourly_forecast",
    "description": "Get hourly weather forecast for a specified location",
    "endpoint": "/tools/accuweather/hourly",
    "method": "POST",
    "input_schema": {
      "lat": "number",
      "lon": "number",
      "hours": "integer",
      "units": "string",
      "lang": "string"
    }
  },
  {
    "name": "accuweather_daily_forecast",
    "description": "Get daily weather forecast for a specified location",
    "endpoint": "/tools/accuweather/daily",
    "method": "POST",
    "input_schema": {
      "locationKey": "string",
      "days": "integer",
      "units": "string",
      "lang": "string"
    }
  }
]

Use Cases

• Chatbot weather queries: A conversational assistant can call the hourly tool to answer “What’s the weather like for