Java基于Spring AI开发MCP

发表于 分类于
MCP是一种标准化协议,使AI模型能够以结构化的方式与外部工具和资源进行交互.本文将介绍如何基于SpringAI开发MCP服务端和客户端.

文章摘要

  • MCP 和 Spring AI相关文档
  • 开发一个Spring天气应用,并发布为MCP Server.
  • 开发一个Spring应用作为MCP Client调用自己的MCP Server服务,并获取天气信息.

什么是MCP?

文档: https://modelcontextprotocol.io/

MCP 是一种开放协议，它标准化了应用程序如何向 LLM 提供上下文。将 MCP 想象成用于 AI 应用程序的 USB-C 端口。正如 USB-C 提供了一种将设备连接到各种外围设备和配件的标准化方式一样，MCP 也提供了一种将 AI 模型连接到不同数据源(例如本地文件、数据库或内容存储库)和工具(例如 GitHub、Google Maps 或 Puppeteer)的标准化方式。

Spring AI MCP

文档: https://docs.spring.io/spring-ai/reference/api/mcp/mcp-overview.html

构建服务端MCP Server

引入依赖

1
2
3
4
5
6
7
8
9
10
11
<dependencyManagement>
    <dependencies>
        <dependency>
            <groupId>org.springframework.ai</groupId>
            <artifactId>spring-ai-bom</artifactId>
            <version>1.0.0</version>
            <type>pom</type>
            <scope>import</scope>
        </dependency>
    </dependencies>
</dependencyManagement>
1
2
3
4
<dependency>
    <groupId>org.springframework.ai</groupId>
    <artifactId>spring-ai-starter-mcp-server-webmvc</artifactId>
</dependency>

application.yml配置

1
2
3
4
5
6
7
8
9
10
11
12
spring:
  application:
    name: mcp-server
  ai:
    mcp:
      server:
        enabled: true
        name: my-weather-and-ip-mcp-server
        type: SYNC
        instructions: 这里是mcp介绍
server:
  port: 8888

编写工具Tools

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
package org.example.mcpserver;

import com.fasterxml.jackson.core.JsonProcessingException;
import com.fasterxml.jackson.databind.ObjectMapper;
import org.springframework.ai.tool.annotation.Tool;
import org.springframework.ai.tool.annotation.ToolParam;
import org.springframework.stereotype.Service;
import org.springframework.web.client.RestClient;

import java.net.URI;
import java.util.Map;

@Service
public class WeatherService {

    @Tool(description = "根据城市名获取天气信息")
    String getCurrentWeather(@ToolParam(description = "城市名") String cityName) {
        System.err.printf("准备查询【%s】天气预报%n", cityName);
        RestClient client = RestClient.create(URI.create("https://api.vvhan.com"));
        Map<?, ?> result = client.get()
                .uri("/api/weather?city={0}", cityName)
                .retrieve()
                .body(Map.class);
        try {
            return new ObjectMapper().writeValueAsString(result);
        } catch (JsonProcessingException e) {
            throw new RuntimeException(e);
        }
    }

}

配置Tools

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
package org.example.mcpserver;

import org.springframework.ai.tool.ToolCallbackProvider;
import org.springframework.ai.tool.method.MethodToolCallbackProvider;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;

@Configuration
public class ToolsConfig {

    @Bean
    public ToolCallbackProvider weatherTools(WeatherService weatherService) {
        return MethodToolCallbackProvider.builder().toolObjects(weatherService).build();
    }

}

构建客户端MCP Client

我连接的模型是本地Ollama部署的qwen3:8b,因此需引入依赖spring-ai-starter-model-ollama

文章: 使用Ollama本地部署DeepSeek-R1大模型

引入依赖

1
2
3
4
5
6
7
8
9
10
11
<dependencyManagement>
    <dependencies>
        <dependency>
            <groupId>org.springframework.ai</groupId>
            <artifactId>spring-ai-bom</artifactId>
            <version>1.0.0</version>
            <type>pom</type>
            <scope>import</scope>
        </dependency>
    </dependencies>
</dependencyManagement>
1
2
3
4
5
6
7
8
9
10
<dependency>
    <groupId>org.springframework.ai</groupId>
    <artifactId>spring-ai-starter-model-ollama</artifactId>
</dependency>

<dependency>
    <groupId>org.springframework.ai</groupId>
    <artifactId>spring-ai-starter-mcp-client</artifactId>
</dependency>

application.yml配置

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
spring:
  application:
    name: mcp-client
  ai:
    ollama:
      base-url: http://localhost:11434
      chat:
        options:
           #  注: deepseek不支持MCP,改用qwen3
          model: qwen3:8b
    mcp:
      client:
        enable: true
        name: spring-ai-mcp-client
        type: SYNC
        sse:
          connections:
            server1:
              url: http://localhost:8888
              sse-endpoint: /sse
logging:
  level:
    org.springframework.ai.chat.client.advisor: debug

配置自定义ChatClient

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
package org.example.mcpclient.config;

import org.springframework.ai.chat.client.ChatClient;
import org.springframework.ai.chat.client.advisor.SimpleLoggerAdvisor;
import org.springframework.ai.tool.ToolCallbackProvider;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;

@Configuration
public class AiConfig {

    @Bean
    public ChatClient chatClient(ChatClient.Builder chatClientBuilder, ToolCallbackProvider mcpTools) {
        return chatClientBuilder
                .defaultSystem("你是一名机器人助手,名字叫小花花")
                .defaultAdvisors(new SimpleLoggerAdvisor())
                // 注入MCP工具
                .defaultTools(mcpTools)
                .build();
    }

}

Controller

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
package org.example.mcpclient.controller;

import io.modelcontextprotocol.client.McpSyncClient;
import lombok.RequiredArgsConstructor;
import lombok.extern.slf4j.Slf4j;
import org.springframework.ai.chat.client.ChatClient;
import org.springframework.ai.mcp.SyncMcpToolCallbackProvider;
import org.springframework.ai.tool.ToolCallback;
import org.springframework.web.bind.annotation.GetMapping;
import org.springframework.web.bind.annotation.RequestMapping;
import org.springframework.web.bind.annotation.RestController;

import java.util.List;

@Slf4j
@RequiredArgsConstructor
@RestController
@RequestMapping("/ai")
class ChatController {

    private final ChatClient chatClient;
    private final List<McpSyncClient> mcpSyncClients;
    private final SyncMcpToolCallbackProvider toolCallbackProvider;

    /**
     * prompt: 今天深圳天气怎么样?
     * @param prompt
     * @return
     */
    @GetMapping("/chat")
    public String chat(String prompt) {
        return chatClient.prompt(prompt)
                .call().content();
    }

    /**
     * 打印mcp信息
     * MCP工具参考: https://docs.spring.io/spring-ai/reference/api/mcp/mcp-helpers.html
     * @param prompt
     * @return
     */
    @GetMapping("/printMcp")
    public String printMcp(String prompt) {
        for (McpSyncClient mcpSyncClient : mcpSyncClients) {
            log.info(String.valueOf(mcpSyncClient.getServerInfo()));
            log.info(mcpSyncClient.getServerInstructions());
        }
        for (ToolCallback toolCallback : toolCallbackProvider.getToolCallbacks()) {
            log.info(String.valueOf(toolCallback.getToolDefinition()));
            log.info(toolCallback.call("{\"cityName\": \"深圳\"}"));
        }
        return "SUCCESS";
    }

}

测试

启动程序后访问接口: http://localhost:8080/ai/chat?prompt=明天深圳天气怎么样?

问答测试