A comprehensive Model Context Protocol (MCP) server built with Spring AI and Spring Boot, providing a solid foundation for building MCP-enabled applications. This server supports both STDIO transport (for Claude Desktop integration) and SSE transport (for web-based clients).

For more information, see the MCP Server Boot Starter reference documentation.

This MCP server demonstrates:

• Integration with spring-ai-mcp-server-webflux-spring-boot-starter
• Dual transport support: STDIO and SSE (Server-Sent Events)
• Automatic tool registration using Spring AI's @Tool annotation
• Clean separation of concerns with dedicated ToolsService
• Production-ready logging configuration
• Comprehensive test coverage with 27+ unit tests

• Description: Capitalize the first letter of each word in the input text
• Parameters:
  • text (String): Input text to capitalize
• Example: "hello world" → "Hello World"

• Description: Perform basic mathematical operations
• Parameters:
  • number1 (double): First number
  • number2 (double): Second number
  • operator (String): Mathematical operator (+, -, *, /, %, ^)
• Example: calculate(15, 3, "+") → 18.0

• Java 21+
• Maven 3.6+

./mvnw clean install

java -Dspring.profiles.active=stdio -jar target/mcp-server-0.0.1-SNAPSHOT.jar

java -jar target/mcp-server-0.0.1-SNAPSHOT.jar

Server will be available at http://localhost:8080/mcp/message

Use the included test script:

# Test STDIO mode
./test-mcp.sh --stdio --test-tools

# Test SSE mode  
./test-mcp.sh --sse --test-tools

# Build and test both modes
./test-mcp.sh --build --both --test-tools

Add this to your Claude Desktop claude_desktop_config.json:

{
  "mcpServers": {
    "mcp-server": {
      "command": "java",
      "args": [
        "-Dspring.profiles.active=stdio",
        "-jar",
        "/absolute/path/to/mcp-server/target/mcp-server-0.0.1-SNAPSHOT.jar"
      ]
    }
  }
}

The tools capitalizeText and calculate will be available for use.

Try asking Claude Desktop:

• "Can you capitalize this text: 'hello world from mcp server'"
• "Calculate 15 + 3 using the calculator tool"

The server uses profile-based configuration:

# STDIO Transport for Claude Desktop
spring.ai.mcp.server.stdio=true
spring.main.web-application-type=none
spring.main.banner-mode=off

# Logging to file only (console interferes with MCP protocol)
logging.level.root=OFF
logging.config=classpath:logback-stdio.xml

# SSE Transport for Web Clients
spring.ai.mcp.server.stdio=false
server.port=8080

# MCP endpoint
spring.ai.mcp.server.sse-message-endpoint=/mcp/message

• McpServerApplication: Main Spring Boot application with tool registration
• ToolsService: Service containing MCP tools with @Tool annotations
• Transport Layer: Automatic Spring AI MCP transport configuration
• Configuration: Profile-based setup for different deployment modes

src/
├── main/java/com/baskettecase/mcpserver/
│   ├── McpServerApplication.java      # Main application
│   └── ToolsService.java              # MCP tools implementation
├── main/resources/
│   ├── application.properties         # Base configuration
│   ├── application-stdio.properties   # STDIO transport config
│   ├── application-sse.properties     # SSE transport config
│   └── logback-stdio.xml              # STDIO logging config
└── test/java/
    ├── ToolsServiceTest.java          # Comprehensive tool tests
    └── org/springframework/ai/mcp/sample/client/
        ├── SampleClient.java          # Example MCP client
        ├── ClientStdio.java           # STDIO client example
        └── ClientSse.java             # SSE client example

Add methods to ToolsService with the @Tool annotation:

@Tool(description = "Your tool description")
public String yourTool(String parameter) {
    // Your implementation
    return "result";
}

./mvnw test

• STDIO mode: Logs go to /tmp/mcp-server-stdio.log
• SSE mode: Logs go to console and ./target/mcp-server-sse.log

var stdioParams = ServerParameters.builder("java")
    .args("-Dspring.profiles.active=stdio", "-jar", "target/mcp-server-0.0.1-SNAPSHOT.jar")
    .build();

var transport = new StdioClientTransport(stdioParams);
var client = McpClient.sync(transport).build();

var transport = new WebFluxSseClientTransport(
    WebClient.builder().baseUrl("http://localhost:8080"));
var client = McpClient.sync(transport).build();

Key dependencies include:

<dependency>
    <groupId>org.springframework.ai</groupId>
    <artifactId>spring-ai-mcp-server-webflux-spring-boot-starter</artifactId>
    <version>1.1.0-SNAPSHOT</version>
</dependency>

This starter provides:

• Reactive and STDIO transport support
• Auto-configured MCP endpoints
• Tool callback registration
• Spring WebFlux integration

This project serves as a foundation for MCP server development. Feel free to:

• Add new tools to ToolsService
• Extend transport configurations
• Improve test coverage
• Add new MCP capabilities

• Spring AI Documentation
• MCP Server Boot Starter Docs
• Model Context Protocol Specification
• Claude Desktop MCP Guide

This project is provided as-is for educational and development purposes.