Spring AI MCP Client + DeepSeek R1 搭建自定义可扩展的AI助手服务

一、Spring AI MCP Client

MCP 由 Anthropic 推出的一种开放标准协议，旨在统一大模型（LLM）与外部数据源和工具之间的通信方式。通过 MCP 协议，开发者可以更高效地实现 AI 模型与外部资源的集成，从而提升应用的智能化和上下文感知能力。

上篇文章介绍了关于 Spring AI MCP Server + Cline 的方式实现了数据库 ChatBi 助手，但MCP客户端的能力是基于 Cline 进行实现的，本篇文章继续进行深入，通过 Spring AI MCP Client 端框架调用MCP Server 端的 Tools，AIGC 能力使用 deepseek-r1 进行完成。

上篇文章在实现 MCP Server 端时采用的 stdio 模式进行的接入，本篇文章也继续扩展重新采用 SSE 模式进行实现，同样是实现一个 ChatBi 数据查询助手 ，不过关于表结构的创建，还是请参考上篇文章中的介绍：

Spring AI MCP Server + Cline 快速搭建一个数据库 ChatBi 助手

关于AIGC能力端的说明，本次使用的 OpenRouter 平台提供的大模型能力，非局限于当前平台你可以使用任何支持 tools 调用的AIGC平台。

Postman测试调用Client端接口效果如下：

Spring AI MCP Client 端的官方介绍文档如下：

https://docs.spring.io/spring-ai/reference/api/mcp/mcp-client-boot-starter-docs.html

二、Spring AI MCP Server SSE 模式搭建

1. 创建父 Maven

创建父 Maven 项目，在 pom 中加入如下全局声明，这里 Java 使用17版本，SpringBoot 选用 3.3.0 版本，Spring AI 选用 1.0.0-SNAPSHOT 版本。

<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
         xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 https://maven.apache.org/xsd/maven-4.0.0.xsd">
    <modelVersion>4.0.0</modelVersion>
    <groupId>com.example</groupId>
    <artifactId>mcp-demo</artifactId>
    <version>0.0.1-SNAPSHOT</version>
    <name>mcp-demo</name>
    <description>mcp-demo</description>
    <packaging>pom</packaging>
    <properties>
        <java.version>17</java.version>
        <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
        <project.reporting.outputEncoding>UTF-8</project.reporting.outputEncoding>
        <spring-boot.version>3.3.0</spring-boot.version>
        <spring-ai.version>1.0.0-SNAPSHOT</spring-ai.version>
    </properties>

    <modules>
        <module>mcp-server</module>
        <module>mcp-client</module>
    </modules>

    <dependencies>
        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter</artifactId>
        </dependency>

        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-test</artifactId>
            <scope>test</scope>
        </dependency>

        <dependency>
            <groupId>org.projectlombok</groupId>
            <artifactId>lombok</artifactId>
        </dependency>
    </dependencies>


    <dependencyManagement>
        <dependencies>
            <dependency>
                <groupId>org.springframework.boot</groupId>
                <artifactId>spring-boot-dependencies</artifactId>
                <version>${spring-boot.version}</version>
                <type>pom</type>
                <scope>import</scope>
            </dependency>
            <dependency>
                <groupId>org.springframework.ai</groupId>
                <artifactId>spring-ai-bom</artifactId>
                <version>${spring-ai.version}</version>
                <type>pom</type>
                <scope>import</scope>
            </dependency>
        </dependencies>
    </dependencyManagement>


    <repositories>
        <repository>
            <name>Central Portal Snapshots</name>
            <id>central-portal-snapshots</id>
            <url>https://central.sonatype.com/repository/maven-snapshots/</url>
            <releases>
                <enabled>false</enabled>
            </releases>
            <snapshots>
                <enabled>true</enabled>
            </snapshots>
        </repository>
        <repository>
            <id>spring-milestones</id>
            <name>Spring Milestones</name>
            <url>https://repo.spring.io/milestone</url>
            <snapshots>
                <enabled>false</enabled>
            </snapshots>
        </repository>
        <repository>
            <id>spring-snapshots</id>
            <name>Spring Snapshots</name>
            <url>https://repo.spring.io/snapshot</url>
            <releases>
                <enabled>false</enabled>
            </releases>
        </repository>
    </repositories>

</project>

2. 创建 mcp-server 子模块

在父依赖创建SpringBoot子模块，修改pom依赖如下：

<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
         xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 https://maven.apache.org/xsd/maven-4.0.0.xsd">
    <parent>
        <groupId>com.example</groupId>
        <artifactId>mcp-demo</artifactId>
        <version>0.0.1-SNAPSHOT</version>
    </parent>
    <modelVersion>4.0.0</modelVersion>
    <artifactId>mcp-server</artifactId>
    <version>0.0.1-SNAPSHOT</version>
    <name>mcp-server</name>
    <description>mcp-server</description>
    <properties>
        <java.version>17</java.version>
        <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
        <project.reporting.outputEncoding>UTF-8</project.reporting.outputEncoding>
    </properties>
    <dependencies>

        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-web</artifactId>
        </dependency>

        <dependency>
            <groupId>org.springframework.ai</groupId>
            <artifactId>spring-ai-mcp-server-webmvc-spring-boot-starter</artifactId>
            <version>${spring-ai.version}</version>
        </dependency>

        <dependency>
            <groupId>mysql</groupId>
            <artifactId>mysql-connector-java</artifactId>
            <version>8.0.28</version>
        </dependency>

        <dependency>
            <groupId>com.alibaba</groupId>
            <artifactId>druid</artifactId>
            <version>1.1.6</version>
        </dependency>

        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-jdbc</artifactId>
        </dependency>
    </dependencies>

    <build>
        <plugins>
            <plugin>
                <groupId>org.apache.maven.plugins</groupId>
                <artifactId>maven-compiler-plugin</artifactId>
                <version>3.8.1</version>
                <configuration>
                    <source>17</source>
                    <target>17</target>
                    <encoding>UTF-8</encoding>
                </configuration>
            </plugin>
            <plugin>
                <groupId>org.springframework.boot</groupId>
                <artifactId>spring-boot-maven-plugin</artifactId>
                <version>${spring-boot.version}</version>
                <executions>
                    <execution>
                        <goals>
                            <goal>repackage</goal>
                        </goals>
                    </execution>
                </executions>
            </plugin>
        </plugins>
    </build>

</project>

application.yml 配置：

server:
  port: 8010

spring:
  datasource:
    url: jdbc:mysql://127.0.0.1:3306/langchain?useUnicode=true&characterEncoding=utf8&serverTimezone=GMT%2B8
    type: com.alibaba.druid.pool.DruidDataSource
    username: root
    password: root
    driver-class-name: com.mysql.cj.jdbc.Driver

  ai:
    mcp:
      server:
        name: mymcp
        type: SYNC
        sse-message-endpoint: /mcp/messages

3. 创建 MCP Tools

创建三个 MCP Tools ，实现 获取可用表名、根据表名获取表结构、执行SQL 三个功能，这点和上篇文章中 stdio 模式是一致的 ：

@Component
public class DBTool {

    @Resource
    private JdbcTemplate jdbcTemplate;

    private final String sql = "SELECT TABLE_NAME, TABLE_COMMENT FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_SCHEMA = 'langchain'";

    private final String schemaSql = "SELECT COLUMN_NAME, DATA_TYPE, COLUMN_COMMENT FROM INFORMATION_SCHEMA.COLUMNS " +
            "WHERE TABLE_SCHEMA = 'langchain' AND TABLE_NAME = ?";

    @Tool(description = "获取所有可用的表名")
    public List<String> getTables() {
        List<Map<String, Object>> maps = jdbcTemplate.queryForList(sql);
        return maps.stream().map(map -> {
            String tableName = String.valueOf(map.get("TABLE_NAME"));
            String tableComment = String.valueOf(map.get("TABLE_COMMENT"));
            return tableName + " COMMENT " + tableComment;
        }).collect(Collectors.toList());
    }

    @Tool(description = "根据表名获取Schema")
    public String getTableSchema(@ToolParam(description = "表名") List<String> tables) {
        return tables.stream().filter(t -> !t.isBlank()).map(tableName -> {
            List<Map<String, Object>> columns = jdbcTemplate.queryForList(schemaSql, tableName);
            String tablePrompt = columns.stream().map(map -> {
                String name = String.valueOf(map.get("COLUMN_NAME"));
                String type = String.valueOf(map.get("DATA_TYPE"));
                String comment = String.valueOf(map.get("COLUMN_COMMENT"));
                return String.format("%s (%s) COMMENT %s", name, type, comment);
            }).collect(Collectors.joining(", \n"));
            return String.format("Table: %s (%s)\n", tableName, tablePrompt);
        }).collect(Collectors.joining("\n"));
    }

    @Tool(description = "执行SQL查询结果")
    public List<Map<String, Object>> runSql(@ToolParam(description = "sql") String sql) {
        try {
            if (sql.contains("DELETE") || sql.contains("UPDATE") || sql.contains("INSERT")) {
                throw new RuntimeException("执行SQL仅限于查询语句！");
            }
            return jdbcTemplate.queryForList(sql);
        } catch (RuntimeException e) {
            return Collections.singletonList(Map.of("执行SQL异常",e.getMessage()));
        }
    }

}

4. 注册 MCP Tools ：

@Configuration
public class MCPConfig {

    @Bean
    public List<ToolCallback> tools(DBTool dbTool) {
        return new java.util.ArrayList<>(List.of(ToolCallbacks.from(dbTool)));
    }
}

5. 启动服务

启动后可以看到日志中打印注册了三个tools：

三、MCP Client 端搭建

1. 创建 mcp-client 子模块

在父依赖创建SpringBoot子模块，修改pom依赖如下：

<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
         xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 https://maven.apache.org/xsd/maven-4.0.0.xsd">
    <parent>
        <groupId>com.example</groupId>
        <artifactId>mcp-demo</artifactId>
        <version>0.0.1-SNAPSHOT</version>
    </parent>
    <modelVersion>4.0.0</modelVersion>
    <artifactId>mcp-client</artifactId>
    <version>0.0.1-SNAPSHOT</version>
    <name>mcp-client</name>
    <description>mcp-client</description>
    <properties>
        <java.version>17</java.version>
        <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
        <project.reporting.outputEncoding>UTF-8</project.reporting.outputEncoding>
    </properties>
    <dependencies>
        
        <dependency>
            <groupId>org.springframework.boot</groupId>
            <artifactId>spring-boot-starter-web</artifactId>
        </dependency>

        <dependency>
            <groupId>org.springframework.ai</groupId>
            <artifactId>spring-ai-mcp-client-spring-boot-starter</artifactId>
            <version>${spring-ai.version}</version>
        </dependency>

        <dependency>
            <groupId>org.springframework.ai</groupId>
            <artifactId>spring-ai-starter-model-openai</artifactId>
            <version>${spring-ai.version}</version>
        </dependency>

    </dependencies>

    <build>
        <plugins>
            <plugin>
                <groupId>org.apache.maven.plugins</groupId>
                <artifactId>maven-compiler-plugin</artifactId>
                <version>3.8.1</version>
                <configuration>
                    <source>17</source>
                    <target>17</target>
                    <encoding>UTF-8</encoding>
                </configuration>
            </plugin>
            <plugin>
                <groupId>org.springframework.boot</groupId>
                <artifactId>spring-boot-maven-plugin</artifactId>
                <version>${spring-boot.version}</version>
                <configuration>
                    <mainClass>com.example.mcpclient.McpClientApplication</mainClass>
                    <skip>true</skip>
                </configuration>
                <executions>
                    <execution>
                        <id>repackage</id>
                        <goals>
                            <goal>repackage</goal>
                        </goals>
                    </execution>
                </executions>
            </plugin>
        </plugins>
    </build>

</project>

application.yml 配置如下，主要声明出大模型的链接和MCPServer端的连接信息。

server:
  port: 8020

spring:
  ai:
    openai:
      base-url: https://openrouter.ai/api/
      api-key: sk-or-v1-xxxxxxxx ## 修改为你的Key
      chat:
        enabled: true
        options:
          model: deepseek/deepseek-r1
    mcp:
      client:
        enabled: true
        name: mymcp
        version: 1.0.0
        type: SYNC
        request-timeout: 300s
        sse:
          connections:
            url: http://localhost:8010

2. 不依赖大模型，使用MCP API 的方式调用SSE远程Tools， 验证连接

首先实验下不依赖大模型，使用 MCP API的方式测试调用远程Tools，验证连接是否正常。

@SpringBootTest
public class MCPSSEClientSpringTest {

    @Resource
    List<McpSyncClient> clients;

    @Test
    public void test() {
        clients.stream()
                .filter(Objects::nonNull)
                .filter(m -> Objects.equals(m.getServerInfo().name(), "mymcp"))
                .findFirst()
                .ifPresent(client -> {
                    // 获取所有的 MCP Tools
                    McpSchema.ListToolsResult toolsList = client.listTools();
                    System.out.println("==============MCP Tools===============");
                    toolsList.tools().forEach(t -> System.out.println(t.toString()));
                    // 调用工具，获取全部的表
                    McpSchema.CallToolResult tables = client.callTool(
                            new McpSchema.CallToolRequest(
                                    "getTables",
                                    Collections.emptyMap()
                            )
                    );
                    System.out.println("==============getTables===============");
                    System.out.println("getTables" + tables.content().get(0));
                    // 调用工具，获取表的Schema
                    McpSchema.CallToolResult tableSchema = client.callTool(
                            new McpSchema.CallToolRequest(
                                    "getTableSchema",
                                    Map.of("arg0", Arrays.asList("user", "role"))
                            )
                    );
                    System.out.println("==============getTableSchema===============");
                    System.out.println("getTableSchema" + tableSchema.content().get(0));
                    // 调用工具，执行SQL
                    McpSchema.CallToolResult runSQL = client.callTool(
                            new McpSchema.CallToolRequest(
                                    "runSql",
                                    Map.of("arg0", "select count(id) from user")
                            )
                    );
                    System.out.println("==============runSql===============");
                    System.out.println("runSql" + runSQL.content().get(0));
                });
    }

}

运行后看到 Tools 列表和调用 Tool 成功即表示连接正常。

3. 结合 DeepSeek 自动 Tools 调用

这里临时使用 Map 存放多轮对话，通过 userId 区分具体用户。关于 Tools 的组装，
SpringAI 已经自动将 MCP Tools 放入了 ToolCallbackProvider 中，因此可直接将该对象提供给 ChatClient 即可。

这部分的内容可参考：

https://docs.spring.io/spring-ai/reference/api/tools.html

完整实现逻辑如下：

@RestController
public class ChatController {

    private final ChatClient chatClient;

    // 存放会话，临时使用Map, 实际使用请考虑持久化
    private final Map<String, List<Message>> history;

    public ChatController(ChatClient.Builder chatClientBuilder, ToolCallbackProvider toolCallbackProvider) {
        this.chatClient = chatClientBuilder
                .defaultTools(toolCallbackProvider)
                .build();
        history = new ConcurrentHashMap<>();
    }


    @GetMapping("/chat")
    public String mcpChat(@RequestParam(name = "userId", required = true) String userId,
                          @RequestParam(name = "msg", required = true) String msg) {
        if (!history.containsKey(userId)) {
            history.put(userId, new ArrayList<>());
            history.get(userId).add(new SystemMessage("You are an AI assistant that helps people find information."));
        }
        history.get(userId).add(new UserMessage(msg));
        String result = chatClient
                .prompt(new Prompt(history.get(userId)))
                .call().content();
        history.get(userId).add(new AssistantMessage(result));
        return result;
    }
}

启动服务，下面进行测试。

四、接口测试