Unlock AI Data Integration with Spring AI’s Model Context Protocol (MCP)

MCP: The "Universal Translator" for AI and the External World

As AI technology rapidly advances, efficiently acquiring and processing external data becomes a key challenge. Spring AI now supports the Model Context Protocol (MCP), opening a new world of data interconnection for AI applications.

Why is MCP special?

Unified Standard : One development, run everywhere. MCP acts as a universal adapter allowing AI apps to converse with any data source in the same way.

Full Compatibility : Handles both structured database tables and unstructured text files.

Plug‑and‑Play : Enables direct viewing of GitHub code or Google Docs within a chat interface.

MCP vs Function Calling: More Than Simple Calls

Function Calling is a one‑off custom task, whereas MCP serves as a general translator that builds bridges between systems, allowing free flow of data and functionality.

Spring AI + MCP: Data Interconnection in a Few Lines

With MCP, AI can directly read local documents, databases, and other sources, greatly expanding practical use cases such as content creation and document processing.

1. Add Dependency

<dependency>
    <groupId>org.springframework.experimental</groupId>
    <artifactId>spring-ai-mcp</artifactId>
    <version>0.4.1</version>
</dependency>

2. Basic Configuration

spring.ai.openai.base-url=your-api-base-url
spring.ai.openai.api-key=your-api-key

3. Core Implementation

@Bean
public CommandLineRunner predefinedQuestions(ChatClient.Builder chatClientBuilder,
   List<McpFunctionCallback> functionCallbacks, ConfigurableApplicationContext context) {
    return args -> {
        var chatClient = chatClientBuilder
            .defaultFunctions(functionCallbacks.toArray(new McpFunctionCallback[0]))
            .build();
        String question1 = "mcp.txt 这篇文档有什么建议吗？方便公众号传播推广";
        System.out.println("ASSISTANT: " + chatClient.prompt(question1).call().content());
        context.close();
    };
}

@Bean
public List<McpFunctionCallback> functionCallbacks(McpSyncClient mcpClient) {
    return mcpClient.listTools(null)
        .tools()
        .stream()
        .map(tool -> new McpFunctionCallback(mcpClient, tool))
        .toList();
}

@Bean(destroyMethod = "close")
public McpSyncClient mcpClient() {
    var stdioParams = ServerParameters.builder("npx")
        .args("-y", "@modelcontextprotocol/server-filesystem", getDbPath())
        .build();
    var mcpClient = McpClient.using(new StdioClientTransport(stdioParams))
        .requestTimeout(Duration.ofSeconds(10)).sync();
    var init = mcpClient.initialize();
    System.out.println("MCP Initialized: " + init);
    return mcpClient;
}

private static String getDbPath() {
    return Paths.get(System.getProperty("user.dir"), "target").toString();
}

Test Execution

Run the following command to start the Spring Boot application and test MCP functionality:

mvn springboot:run