MCP Weather Chat
Model Context Protocol (MCP) integration for interacting with external weather services.
Purpose
Demonstrates capabilities:
- MCP integration for external services
- Weather API integration through protocol
- Tool calling through MCP server
- Extensible architecture for external data
Key Features
- MCP Integration: Model Context Protocol for external tools
- Weather Service: Real-time weather data
- Server Architecture: Separate MCP server
- Tool Calling: Structured function calling for weather
- Extensible Design: Easy addition of new services
Technology Stack
- Runtime: Bun
- Language: TypeScript
- AI Framework: agent-swarm-kit
- Protocol: Model Context Protocol (MCP)
- External APIs: Weather services
Project Structure
src/
├── config/
│ └── mcp.ts # MCP configuration
├── repl.ts # REPL interface
└── lib/
└── swarm.ts # Swarm configuration
server/
└── src/
└── index.ts # MCP server
Installation and Setup
# Install dependencies
bun install
# Start MCP server (terminal 1)
cd server && bun run src/index.ts
# Start client (terminal 2)
bun run src/repl.ts
Configuration
Create a .env file:
WEATHER_API_KEY=your_weather_api_key
MCP_SERVER_URL=http://localhost:3001
OPENAI_API_KEY=your_openai_api_key
Usage Examples
- Current weather: "What's the weather in Moscow?"
- Forecast: "Tomorrow's forecast for St. Petersburg"
- Comparison: "Compare weather in Moscow and London"
- Planning: "Should I take an umbrella in Sochi today?"
MCP Benefits
- Standardized Protocol: Unified standard for external services
- Type Safety: Structured data exchange
- Extensibility: Easy addition of new tools
- Isolation: Separate servers for different services
Use Cases
Excellent for:
- Weather applications
- Travel planning assistants
- Outdoor activity recommendations
- Agricultural advice systems
- Event planning tools
Extension
MCP architecture allows adding:
- News services
- Stock market data
- Translation services
- Calendar integration
- Maps and navigation