will8ug/mcpwithspring-server

MCP Weather Server with Spring WebFlux and Spring AI

A simple Model Context Protocol (MCP) server built with Spring WebFlux and Spring AI that provides weather information through a mocked weather API.

Features

• Weather Service: Mock weather data service that returns temperature, condition, humidity, and wind speed
• Spring AI MCP Tools: Two MCP tools available using Spring AI tool annotations:
  • get_weather: Returns detailed weather data for a city
  • get_weather_summary: Returns a human-readable weather summary for a city
• REST API: Standard REST endpoints for direct weather queries
• Reactive: Built with Spring WebFlux for non-blocking, reactive programming
• Spring AI Integration: Uses Spring AI's MCP server WebFlux starter for seamless MCP integration

Dependencies

• Spring Boot 3.5.4
• Spring WebFlux
• Spring AI MCP Server WebFlux Starter
• Spring AI Model (for tool annotations)
• MCP SDK 0.10.0
• Reactor Test (for testing)

Running the Application

1. Start the server:

   ./mvnw spring-boot:run
   

2. Test the REST API:

   # Get weather data for a city
   curl http://localhost:8080/api/weather/London
   
   # Get weather summary for a city
   curl http://localhost:8080/api/weather/London/summary
   

MCP Tools

The application uses Spring AI's @Tool annotation to automatically register MCP tools:

get_weather

Returns detailed weather information for a specified city.

Parameters:

• city (string, required): The name of the city

Example Response:

{
  "city": "London",
  "temperature": "22°C",
  "condition": "Sunny",
  "humidity": "65%",
  "windSpeed": "10 km/h",
  "description": "Clear skies with light breeze"
}

get_weather_summary

Returns a human-readable weather summary for a specified city.

Parameters:

• city (string, required): The name of the city

Example Response:

The weather in London is Sunny with a temperature of 22°C. Clear skies with light breeze.

Architecture

• WeatherService: Core service with Spring AI @Tool annotations for MCP integration
• WeatherController: REST controller for direct API access
• Spring AI MCP Server: Automatically configured MCP server using Spring AI's WebFlux starter

Code Structure

@Service
public class WeatherService {

    @Tool(name = "get_weather", description = "Get current weather information for a specific city")
    public Mono<Map<String, Object>> getWeather(String city) {
        // Mock weather data implementation
    }

    @Tool(name = "get_weather_summary", description = "Get a human-readable weather summary for a specific city")
    public Mono<String> getWeatherSummary(String city) {
        // Weather summary implementation
    }
}

Testing

Run the tests with:

./mvnw test

The test suite includes:

• Weather service functionality tests
• Weather data structure validation
• Weather summary generation tests

MCP Integration

The Spring AI MCP server WebFlux starter automatically:

• Discovers @Tool annotated methods
• Registers them as MCP tools
• Provides MCP protocol endpoints
• Handles tool execution and response formatting

This approach leverages Spring AI's built-in MCP capabilities for a clean, maintainable implementation.