mcp-mianshiya-server

MCP.Pizza Chef: yuyuanweb

Mianshiya is a Chinese interview-practice site, and this connector does exactly one thing: you describe a topic and it hands back up to five matching question titles with links to read them on mianshiya.com. Despite the project's own summary, no answers come back, only titles and links, so you finish the job in a browser, and some questions there are members-only. Nothing needs paying for on the connector side. Every result is in Chinese. No code changes since July 2025.

Unmaintained · No commits in 13 months.
Coding
Web/Research

Use This MCP server To

Find real interview questions on a topic I am revising Build a reading list before a technical interview See how a subject is usually phrased by interviewers Collect question links to work through later

README

面试鸭 MCP Server

简介

面试鸭 的题目搜索API现已兼容MCP协议,是国内首家兼容MCP协议的面试刷题网站。关于MCP协议,详见MCP官方文档

依赖MCP Java SDK开发,任意支持MCP协议的智能体助手(如ClaudeCursor以及千帆AppBuilder等)都可以快速接入。

以下会给更出详细的适配说明。

工具列表

题目搜索 questionSearch
  • 将面试题目检索为面试鸭里的题目链接
  • 输入: 题目
  • 输出: [题目](链接)

快速开始

使用面试鸭MCP Server主要通过Java SDK 的形式

Java 接入

前提需要Java 17 运行时环境

安装
git clone https://github.com/yuyuanweb/mcp-mianshiya-server
构建
cd mcp-mianshiya-server
mvn clean package
使用
  1. 打开Cherry Studio设置,点击MCP 服务器cherry1.png

  2. 点击编辑 JSON,将以下配置添加到配置文件中。

{
  "mcpServers": {
    "mianshiyaServer": {
      "command": "java",
      "args": [
        "-Dspring.ai.mcp.server.stdio=true",
        "-Dspring.main.web-application-type=none",
        "-Dlogging.pattern.console=",
        "-jar",
        "/yourPath/mcp-server-0.0.1-SNAPSHOT.jar"
      ],
      "env": {}
    }
  }
}

cherry2.png

  1. 在设置-模型服务里选择一个模型,输入API密钥,选择模型设置,勾选下工具函数调用功能。 cherry3.png
  2. 在输入框下面勾选开启MCP服务。 cherry4.png
  3. 配置完成,然后查询下面试题目 cherry5.png
代码调用
  1. 引入依赖
        <dependency>
            <groupId>com.alibaba.cloud.ai</groupId>
            <artifactId>spring-ai-alibaba-starter</artifactId>
            <version>1.0.0-M6.1</version>
        </dependency>
    <dependency>
      <groupId>org.springframework.ai</groupId>
      <artifactId>spring-ai-mcp-client-spring-boot-starter</artifactId>
      <version>1.0.0-M6</version>
    </dependency>
  1. 配置MCP服务器 需要在application.yml中配置MCP服务器的一些参数:
spring:
  ai:
    mcp:
      client:
        stdio:
          # 指定MCP服务器配置文件
          servers-configuration: classpath:/mcp-servers-config.json
  mandatory-file-encoding: UTF-8

其中mcp-servers-config.json的配置如下:

{
  "mcpServers": {
    "mianshiyaServer": {
      "command": "java",
      "args": [
        "-Dspring.ai.mcp.server.stdio=true",
        "-Dspring.main.web-application-type=none",
        "-Dlogging.pattern.console=",
        "-jar",
        "/Users/gulihua/Documents/mcp-server/target/mcp-server-0.0.1-SNAPSHOT.jar"
      ],
      "env": {}
    }
  }
}

客户端我们使用阿里巴巴的通义千问模型,所以引入spring-ai-alibaba-starter依赖,如果你使用的是其他的模型,也可以使用对应的依赖项,比如openAI引入spring-ai-openai-spring-boot-starter 这个依赖就行了。 配置大模型的密钥等信息:

spring:
  ai:
    dashscope:
      api-key: ${通义千问的key}
      chat:
        options:
          model: qwen-max

通义千问的key可以直接去官网 去申请,模型我们用的是通义千问-Max。 3) 初始化聊天客户端

@Bean
public ChatClient initChatClient(ChatClient.Builder chatClientBuilder,
                                 ToolCallbackProvider mcpTools) {
    return chatClientBuilder
    .defaultTools(mcpTools)
    .build();
}
  1. 接口调用
    @PostMapping(value = "/ai/answer/sse", produces = MediaType.TEXT_EVENT_STREAM_VALUE)
    public Flux<String> generateStreamAsString(@RequestBody AskRequest request) {

        Flux<String> content = chatClient.prompt()
                .user(request.getContent())
                .stream()
                .content();
        return content
                .concatWith(Flux.just("[complete]"));

    }

mcp-mianshiya-server FAQ

Does it give me the answers as well as the questions?
No, and this is the main thing the project oversells. The code returns at most five question titles and their web links, nothing else. You open the site to read any answer.
Do I need to pay or sign up for anything?
Not to run the connector, which sends no login details at all. Reading some of the questions it links to does need a paid membership on the site itself.
What language are the results in?
Chinese, throughout. The site, the questions and the answers are all Chinese-language, so it is only useful if you read Chinese.
Can I narrow results by difficulty or subject tag?
No. The request it builds has slots for difficulty, tags and whether an answer exists, but the code never fills any of them in. Only your search words are sent.
Which apps does it work in?
Anything that can launch a local command, including Cherry Studio, Claude Desktop and Cursor. The project's own walkthrough uses Cherry Studio.
How hard is setup?
Harder than the description suggests. There is no ready-made package: you install Java 17, download the code, build it yourself, then point your app at the file that comes out.
Is it still maintained?
The last change was July 2025, so it has been quiet for over a year. It leans on the site's own search staying available.