MCP Weather Server
A production-grade Model Context Protocol server that provides real-time weather data and 5-day forecasts using the Open-Meteo API. It offers comprehensive meteorological metrics supported by Redis caching and built-in performance monitoring tools.
README
🌤️ MCP Weather Server
A production-grade Model Context Protocol (MCP) server providing comprehensive real-time weather data using the free Open-Meteo API.
✨ Features
🌦️ Weather Data
- Current Weather: Real-time conditions with 20+ meteorological parameters
- 5-Day Forecast: Detailed daily predictions with solar, UV, and precipitation data
- Comprehensive Metrics: Temperature, humidity, wind, pressure, UV index, visibility, cloud cover
- Solar Data: Sunshine duration, daylight hours, solar radiation, evapotranspiration
🚀 Performance & Reliability
- Redis Caching: Intelligent caching with configurable TTL (15min current, 1h forecast)
- Metrics & Monitoring: Real-time performance tracking and error logging
- Graceful Degradation: Continues working without Redis/cache if unavailable
- Type-Safe: Full Zod validation for all inputs and API responses
🏗️ Architecture
- Multiple Transports: Stdio (MCP Inspector) + HTTP/SSE streaming
- Factory Pattern: Isolated server instances per session
- Layered Design: Services, Tools, Server separation
- ESM Modules: Modern JavaScript with full ES2022 support
🔧 Developer Experience
- 64 Unit Tests: Comprehensive Jest test suite
- Hot Reload: Development mode with file watching
- Debug Support: Integrated debugging for both transports
- Rich Logging: Structured logs with performance metrics
📊 Data Sources
Open-Meteo API (Free Tier)
- No API Key Required - Unlimited calls for non-commercial use
- Global Coverage - Weather models for every location worldwide
- Multiple Models - ECMWF, NOAA GFS, Météo-France, DWD ICON, and more
- Real-time Updates - 15-minute to hourly resolution
Available Weather Variables
Current Conditions
- Temperature & "Feels Like" (2m & apparent)
- Humidity, Dew Point, Vapour Pressure Deficit
- Wind Speed/Direction/Gusts (10m, 80m, 120m, 180m altitudes)
- Pressure (MSL & Surface), Cloud Cover (Total/Low/Mid/High)
- Visibility, UV Index, Precipitation
- Evapotranspiration & Reference ET₀
Forecast Data
- Daily min/max temperatures (actual & apparent)
- Precipitation (total, rain, showers, snow)
- Wind (speed, gusts, dominant direction)
- Solar (sunrise/sunset, sunshine/daylight duration, UV index)
- Radiation (shortwave solar radiation sum)
🏛️ Architecture
src/
├── index.js # Entry point & transport orchestration
├── server/
│ ├── weatherMcpFactory.js # Server factory with cache init
│ └── streamableHttpServer.js # HTTP/SSE transport server
├── services/
│ ├── weatherApi.js # Open-Meteo API client with caching
│ ├── weatherCodes.js # WMO weather code mappings
│ ├── cache.js # Redis cache service
│ └── metrics.js # Performance monitoring
└── tools/
├── schemas.js # Zod validation schemas
├── getWeatherTool.js # Current weather MCP tool
├── getWeatherForecast5DaysTool.js # Forecast MCP tool
└── getServerStatsTool.js # Server metrics tool
Design Patterns
| Pattern | Implementation | Purpose |
|---|---|---|
| Factory | createWeatherMcpServer() |
Isolated server instances per session |
| Strategy | Stdio vs HTTP transport strategies | Multiple deployment modes |
| Adapter | MCP SDK transport wrapping | Unified interface for different transports |
| Decorator | Response formatting with emojis | Enhanced user experience |
| Observer | Metrics collection | Performance monitoring |
| Proxy | Cache layer | API call optimization |
📋 Installation
Prerequisites
- Node.js ≥16.0.0 (with
--experimental-vm-modulesfor Jest) - Redis (optional, for caching) - Install from redis.io
Setup
# Clone repository
git clone <repository-url>
cd mcp-weather
# Install dependencies
npm install
# Run tests
npm test
# Start server (stdio mode)
npm start
Environment Variables
# Redis configuration (optional)
REDIS_URL=redis://localhost:6379
# HTTP server configuration
MCP_HTTP_PORT=3333
MCP_HTTP_HOST=127.0.0.1
🚀 Usage
MCP Inspector (Recommended)
# Start in stdio mode for MCP Inspector
npm start
# Then open MCP Inspector and connect to the server
HTTP/SSE Mode
# Start HTTP server with SSE transport
npm run start:http
# Server runs on http://127.0.0.1:3333
# MCP endpoint: http://127.0.0.1:3333/mcp
Development Mode
# With hot reload
npm run dev
# With debugging
npm run debug:http # HTTP mode
npm run debug # Stdio mode
🔧 API Reference
Tools
get_weather
Get current weather conditions for coordinates.
Input:
{
"latitude": 48.8566,
"longitude": 2.3522
}
Output:
🌤️ Current Weather
📍 Location: 48.8566, 2.3522
🌡️ Temperature: 22.5°C (feels like 24.1°C)
💧 Humidity: 65%
🌫️ Dew Point: 15.8°C
💨 Wind: 12 km/h from WSW
🌪️ Wind Gusts: 18 km/h
☁️ Conditions: Partly cloudy
☁️ Cloud Cover: 45% (Low: 20%, Mid: 15%, High: 10%)
👁️ Visibility: 24 km
🧭 Pressure: 1013 hPa (MSL) / 1011 hPa (Surface)
☀️ UV Index: 6
💧 Evapotranspiration: 0.15 mm/h
🌱 Reference ET₀: 0.22 mm/h
🌡️ Vapor Pressure Deficit: 0.85 kPa
🌍 Timezone: Europe/Paris
🕐 Updated: 2024-03-29T14:30:00Z
get_weather_forecast
Get 5-day weather forecast for coordinates.
Input:
{
"latitude": 48.8566,
"longitude": 2.3522
}
Output:
📅 5-Day Forecast
📍 Coordinates : 48.8566, 2.3522
🌍 Timezone : Europe/Paris
📅 2024-03-29
🌡️ 18°C – 24°C (feels 16°C – 26°C)
☁️ Partly cloudy
💧 Precipitation: 0.2 mm (0 rain + 0.1 showers + 0 snow)
🌧️ Chance: 15% over 0.5h
🌅 Sunrise: 06:45 | Sunset: 19:32
☀️ Sunshine: 8.2h / Daylight: 12.8h
☀️ UV Index: 6 (Clear sky: 7)
💨 Wind: 15 km/h gusts to 22 km/h from W
⚡ Solar Radiation: 18.5 MJ/m²
🌱 Reference ET₀: 4.2 mm
get_server_stats
Get server performance metrics and cache statistics.
Input: (no parameters required)
Output:
📊 Server Statistics
⏱️ Uptime: 45 minutes
🔄 API Calls: 127
💾 Cache Hits: 89 (70.1%)
❌ Errors: 2
⚡ Avg Response Time: 245ms
🔧 Tool Usage:
• get_weather: 95 calls
• get_weather_forecast: 32 calls
📋 Cache Status:
• Connected: ✅
• Keys: 23
• Memory: 2.4MB
🧪 Testing
Run Test Suite
# Run all tests
npm test
# Run with coverage
npm run test:coverage
# Run in watch mode
npm run test:watch
Test Architecture
- 64 unit tests covering all services and server components
- Mocked API calls for reliable testing without network dependencies
- Cache testing with Redis mock
- Schema validation testing with Zod
- Error handling coverage
Test Files
tests/
├── services/
│ ├── weatherApi.test.js # API client tests
│ └── weatherCodes.test.js # Weather code mapping tests
└── server/
└── weatherMcpFactory.test.js # Server factory tests
📈 Monitoring & Metrics
Built-in Metrics
- API Call Tracking: Count, response times, cache hit rates
- Tool Usage: Per-tool invocation counts
- Error Tracking: Error types and frequencies
- Cache Performance: Hit rates, memory usage, connection status
- Server Health: Uptime, active sessions
Logs
Structured logging with:
- Request/response timing
- Cache operations
- API call details
- Error context
- Performance metrics
Cache Configuration
// Current weather: 15 minutes TTL
// Forecasts: 1 hour TTL
// Automatic expiration and cleanup
🔒 Security & Reliability
Data Validation
- Zod schemas for all inputs
- Coordinate bounds checking (-90/+90, -180/+180)
- API response validation
- Error sanitization (no sensitive data leakage)
Error Handling
- Graceful degradation (continues without cache)
- Structured error responses with appropriate HTTP codes
- Timeout handling for API calls
- Connection retry logic for Redis
Performance
- Connection pooling for Redis
- Request deduplication via caching
- Memory-efficient data structures
- Background cleanup of expired cache entries
🤝 Contributing
Development Setup
# Install dependencies
npm install
# Run tests
npm test
# Start development server
npm run dev
# Debug mode
npm run debug
Code Quality
- ESLint configuration for code style
- Prettier for consistent formatting
- Jest for comprehensive testing
- TypeScript definitions for better IDE support
Architecture Guidelines
- Layer separation: Services, Tools, Server layers
- Dependency injection: Injectable fetch/cache for testing
- Error boundaries: Comprehensive error handling
- Logging standards: Structured logging throughout
📄 License
MIT License - Free for personal and commercial use.
Open-Meteo Terms
- Non-commercial use: Unlimited API calls
- Commercial use: Contact Open-Meteo for licensing
- Attribution required for public deployments
🙏 Acknowledgments
- Open-Meteo for providing free, high-quality weather data
- Anthropic for the Model Context Protocol specification
- Weather model providers: ECMWF, NOAA, Météo-France, DWD, and others
📞 Support
Issues & Bugs
- GitHub Issues: Report bugs and request features
- Documentation: Check this README first
Community
- MCP Discord: Join the Model Context Protocol community
- Open-Meteo Forum: Weather API discussions
Built with ❤️ using the Model Context Protocol
Stdio Mode (Default)
npm start
# or: node src/index.js
Perfect for direct MCP client integration (Claude Desktop, IDEs, etc.)
HTTP/SSE Mode
# Via CLI flag
npm run start:http
# or: node src/index.js --http
# Via environment variables
TRANSPORT=http PORT=8080 node src/index.js
# Default HTTP port: 3001
HTTP Endpoints:
GET /health- Health check endpointGET /sse- Server-Sent Events streamPOST /messages- Tool request endpoint
Demo Script
node demo.js
Shows real-world example calls to weather service for multiple cities.
Run Tests
npm test
Validates input schemas and weather code mappings.
🛠️ Available Tools
Tool: get_weather
Get current weather for a location.
Parameters:
{
"latitude": -90 to 90,
"longitude": -180 to 180
}
Example:
# Request current weather for Paris
{
"latitude": 48.8566,
"longitude": 2.3522
}
Response Example:
☀️ **CURRENT WEATHER**
📍 Location: 48.8566°N, 2.3522°E
🌡️ Temperature: 15°C
☁️ Condition: Partly cloudy
💧 Humidity: 65%
💨 Wind Speed: 12 km/h
🕐 Time: 2024-03-29T10:30:00Z
🌍 Timezone: Europe/Paris
Tool: get_forecast
Get 5-day weather forecast.
Parameters:
{
"latitude": -90 to 90,
"longitude": -180 to 180
}
Response Example:
📅 **5-DAY WEATHER FORECAST**
📍 Location: 48.8566°N, 2.3522°E
🌍 Timezone: Europe/Paris
**Day 1** - 2024-03-29
☁️ Partly cloudy
🔺 Max: 18°C | 🔻 Min: 12°C
**Day 2** - 2024-03-30
🌧️ Slight rain
🔺 Max: 16°C | 🔻 Min: 10°C
...
🌍 Popular Coordinates
| City | Latitude | Longitude |
|---|---|---|
| Paris | 48.8566 | 2.3522 |
| London | 51.5074 | -0.1278 |
| New York | 40.7128 | -74.0060 |
| Tokyo | 35.6762 | 139.6503 |
| Sydney | -33.8688 | 151.2093 |
| Berlin | 52.5200 | 13.4050 |
| Los Angeles | 34.0522 | -118.2437 |
🔗 Data Source
- API: Open-Meteo - Free, open-source weather API
- Documentation: https://open-meteo.com/en/docs
- Weather Codes: WMO Weather Interpretation Codes
📦 Module System
Uses ES Modules (ESM) with "type": "module" for modern JavaScript features:
- Clean import/export syntax
- Top-level await support
- Better tree-shaking compatibility
🧪 Testing
npm test
Test Coverage:
- Input coordinate validation (bounds checking)
- Weather code mapping (WMO codes)
- Schema validation with Zod
- Edge cases and error handling
📝 Project Structure
MCP-Weather/
├── src/
│ ├── index.js # Main entry point
│ ├── server.js # MCP server factory
│ ├── weatherService.js # Weather API service
│ └── weatherTools.js # MCP tools
├── package.json
├── demo.js # Demo script
├── test.js # Test suite
├── README.md
└── .env.example
🔧 Environment Variables
| Variable | Values | Default | Purpose |
|---|---|---|---|
TRANSPORT |
stdio, http |
stdio |
Transport mode |
PORT |
1-65535 |
3001 |
HTTP server port (HTTP mode only) |
Example:
# Run in HTTP mode on custom port
TRANSPORT=http PORT=8080 npm start
🐛 Troubleshooting
| Issue | Solution |
|---|---|
MODULE_NOT_FOUND |
Ensure "type": "module" in package.json |
| Port 3001 in use | Set custom port: PORT=3002 npm run start:http |
| Invalid coordinates error | Check bounds: latitude [-90, 90], longitude [-180, 180] |
| "Unknown weather condition" | WMO code mapping - check Open-Meteo documentation |
| ESM import errors | Verify all imports have .js file extension |
📜 License
ISC
🤝 Contributing
This project follows the MCP specification and best practices for Node.js server development.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。