Skip to main content
Glama
yan5236

bing-cn-mcp-server

by yan5236
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Remote

必应中文搜索 MCP 服务器

让 AI 助手（如 Claude）能够使用必应搜索引擎实时获取网络信息的工具。

什么是 MCP 服务器？

MCP（Model Context Protocol）是一个让 AI 助手能够调用外部工具的协议。这个项目提供了一个必应搜索工具，让 AI 可以帮你搜索网络信息并返回结果。

Related MCP server: Spider MCP

功能特性

  • 🔍 实时搜索: 使用必应中文搜索引擎获取最新网络信息

  • 📄 网页抓取: 自动抓取并提取搜索结果中的网页内容

  • 🌐 中文优化: 专为中文搜索优化，支持简体中文结果

  • 🔒 无需密钥: 不需要申请 API 密钥，开箱即用

  • 🚫 智能过滤: 自动过滤无法抓取的网站（知乎、微信公众号等）

快速开始

在 Claude Desktop 中使用（推荐）

步骤 1: 打开配置文件

根据你的操作系统，找到并打开 Claude Desktop 的配置文件：

Windows 用户:

%AppData%\Claude\claude_desktop_config.json

直接复制上面的路径到文件资源管理器地址栏，然后用记事本打开 claude_desktop_config.json 文件。

macOS 用户:

~/Library/Application Support/Claude/claude_desktop_config.json

在访达中按 Cmd + Shift + G，粘贴上面的路径，然后用文本编辑器打开。

步骤 2: 添加配置

在配置文件中添加以下内容（如果文件是空的，直接粘贴；如果已有内容，在 mcpServers 部分添加）：

{
  "mcpServers": {
    "bing-search": {
      "command": "npx",
      "args": [
        "-y",
        "bing-cn-mcp"
      ]
    }
  }
}

注意，windows下，你应该这样配置：

{
  "mcpServers": {
    "bing-search": {
      "command": "cmd",
      "args": [
        "/c",
        "npx",
        "-y",
        "bing-cn-mcp"
      ]
    }
  }
}

如果配置文件中已经有其他 MCP 服务器，应该像这样：

{
  "mcpServers": {
    "existing-server": {
      "command": "...",
      "args": ["..."]
    },
    "bing-search": {
      "command": "npx",
      "args": [
        "-y",
        "bing-cn-mcp"
      ]
    }
  }
}

步骤 3: 重启 Claude Desktop

完全退出 Claude Desktop（不是最小化，要完全关闭），然后重新打开。

步骤 4: 开始使用

在 Claude 中，你可以这样提问：

"帮我搜索一下人工智能的最新进展"

"搜索 Python 异步编程教程"

Claude 会自动调用必应搜索工具并返回结果。

工具说明

这个 MCP 服务器提供了两个工具：

1. bing_search - 必应搜索

使用必应搜索引擎搜索网络信息。

参数说明:

  • query (必填): 搜索关键词，例如 "人工智能"

  • count (可选): 返回多少条结果，默认 10 条，最多 50 条

  • offset (可选): 从第几条结果开始，用于翻页，默认 0

返回内容: 搜索结果包括：

  • 搜索到的总结果数量

  • 每条结果的标题、链接和摘要

  • 格式化的易读内容

2. crawl_webpage - 网页抓取

抓取并提取网页的文本内容（自动跳过无法访问的网站）。

参数说明:

  • uuids (必填): 搜索结果的 UUID 列表，可以是单个或多个 UUID

  • urlMap (必填): UUID 到 URL 的映射对象，格式: {"uuid1": "url1", "uuid2": "url2"}

返回内容:

  • 网页标题

  • 清理后的正文内容（自动去除广告、导航栏等无关内容）

黑名单网站（这些网站会被自动跳过）:

  • 知乎 (zhihu.com)

  • 小红书 (xiaohongshu.com)

  • 微博 (weibo.com)

  • 微信公众号 (weixin.qq.com)

  • 抖音/TikTok (douyin.com, tiktok.com)

  • B站 (bilibili.com)

  • CSDN (csdn.net)

使用示例

示例 1: 基础搜索

提问: "搜索一下 TypeScript 教程"

AI 会调用: bing_search，参数为 { "query": "TypeScript 教程" }

返回结果:

搜索关键词: TypeScript 教程
找到约 1,234,567 条结果

返回前 10 条结果:
================================================================================

[1] TypeScript 中文手册 - 官方文档
    链接: https://www.tslang.cn/docs/handbook/basic-types.html
    摘要: TypeScript 是 JavaScript 的超集，为其添加了类型系统...

[2] TypeScript 入门教程 - 阮一峰
    链接: https://ts.xcatliu.com/
    摘要: 从 JavaScript 程序员的角度总结思考，循序渐进讲解 TypeScript...

...

示例 2: 自定义结果数量

提问: "搜索 Node.js 性能优化，给我 20 条结果"

AI 会调用: bing_search，参数为 { "query": "Node.js 性能优化", "count": 20 }

示例 3: 翻页搜索

提问: "搜索 React Hooks 教程，从第 11 条结果开始显示"

AI 会调用: bing_search，参数为 { "query": "React Hooks 教程", "offset": 10 }

示例 4: 抓取搜索结果中的网页内容

提问: "搜索 React 教程，并抓取第一个搜索结果的完整内容"

AI 会调用: 先调用 bing_search 搜索，从返回结果中取第一条的 uuid 和 url，再调用 crawl_webpage，参数为 { "uuids": ["<搜索结果中的uuid>"], "urlMap": { "<uuid>": "<对应网址>" } }

注意: crawl_webpage 只能抓取 bing_search 返回的搜索结果，不能直接传入任意网址。

技术实现

本项目使用以下技术栈构建：

  • MCP SDK (@modelcontextprotocol/sdk): 实现 MCP 协议

  • Axios: 发送 HTTP 请求

  • Cheerio: 解析 HTML 页面

  • Zod: 参数验证

  • TypeScript: 类型安全开发

项目结构

bingcnmcp/
├── src/
│   ├── index.ts         # MCP 服务器入口
│   ├── bingSearch.ts    # 必应搜索实现
│   ├── crawler.ts       # 网页抓取实现
│   ├── parser.ts        # HTML 解析器
│   ├── blacklist.ts     # 黑名单配置
│   └── types.ts         # 类型定义
├── build/               # 编译输出
├── package.json
└── tsconfig.json

常见问题

为什么搜索结果是空的？

  1. 网络问题: 检查是否能正常访问 cn.bing.com

  2. 关键词问题: 尝试换个关键词或更具体的搜索词

  3. 被限制: 短时间内搜索太多次可能被必应限制，等待几分钟再试

Claude 没有调用搜索工具怎么办？

  1. 确保配置文件格式正确（JSON 格式）

  2. 确保完全重启了 Claude Desktop

  3. 尝试明确要求 Claude 使用搜索："使用必应搜索工具查找..."

某些网站无法抓取内容

这是正常的，部分网站（如知乎、小红书等）因为访问限制或反爬虫机制被加入了黑名单。搜索结果中会显示这些网站的链接，但不会自动抓取内容。

Windows 系统提示找不到 npx 命令

确保已经安装了 Node.js（推荐 18 或更高版本）。安装后重启电脑或命令行窗口。

下载 Node.js: https://nodejs.org/

注意事项

  1. 网络要求: 需要能够访问 cn.bing.com

  2. 使用频率: 请勿过于频繁搜索，建议每次搜索间隔至少 1-2 秒

  3. 隐私保护: 搜索请求直接发送到必应，本服务器不存储任何搜索记录

  4. 无状态设计: 每次请求都是独立的，不保存 Cookie 或会话信息

  5. 仅供参考：本项目仅供学习参考，请勿用于非法用途。

  6. 结果不对：搜索结果出现不对时，可能触发了反爬，如有更稳定需要，更推荐tavily-mcp

许可证

MIT

贡献

欢迎提交问题报告和改进建议！

Available Tools

2 tools
crawl_webpageA

根据搜索结果的UUID抓取网页内容。支持批量抓取多个网页。会自动过滤黑名单中的网站(如知乎、小红书等)。

ParametersJSON Schema
NameRequiredDescriptionDefault
uuidsYes搜索结果的UUID列表，可以是单个或多个UUID
urlMapYesUUID到URL的映射对象，格式: {"uuid1": "url1", "uuid2": "url2"}

TDQS

A3.9/5.0
Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description must disclose behaviors. It mentions automatic blacklist filtering, which is valuable, but does not state that the operation is read-only or any other behavioral traits like rate limits or authentication needs.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is concise with two sentences, no redundant information. It could be more structured (e.g., bullet points), but it is efficient and front-loaded.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the simplicity of the tool and complete schema coverage, the description adequately covers the essential context: batch crawl, blacklist, and UUID-based input. No output schema exists, so return value description is not required but would be a nice addition.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The input schema covers both parameters with descriptions, and the description adds 'batch support' which aligns with the schema. However, it does not add new meaning beyond what the schema provides, so baseline score of 3 is appropriate.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's purpose: crawling webpage content using UUIDs from search results, with batch support. It differentiates from sibling bing_search by focusing on crawling after search.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description implies usage after obtaining search result UUIDs and mentions blacklist filtering (knowing when not to use for certain sites). However, it lacks explicit guidance on when to use vs alternatives.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

  1. 2 tool updatesv1.0.2
    • First observedbing_search
    • First observedcrawl_webpage
Deploy Server
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

A3.7/5.0

Scored across 2 tools

Disambiguation5/5

The two tools have clearly distinct purposes: bing_search performs searches, and crawl_webpage retrieves web content based on search result UUIDs. There is no overlap.

Naming Consistency4/5

Both tools follow a verb_noun pattern (bing_search, crawl_webpage). The minor inconsistency is that bing_search includes a service prefix (bing) while crawl_webpage does not, but the structure is still uniform.

Tool Count4/5

With only two tools, the set is minimal but well-scoped for its purpose of searching and crawling. It is slightly thin but not unreasonable for a focused utility.

Completeness4/5

The tools cover the essential workflow: search then retrieve full content. There are no obvious gaps for basic use, though advanced features like pagination or filtering are missing.

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

  • Spider MCP

    yc9yc
    A
    license
    B
    quality
    D
    maintenance
    Enables web searching and webpage scraping using pure crawler technology without requiring official APIs. Supports Bing web and news search, batch webpage scraping, and content extraction through Puppeteer automation.
    4
    1
    MIT

  • bingcn

    xiaobenyang-com
    A
    license
    B
    quality
    D
    maintenance
    A Chinese Bing search tool based on MCP protocol, which supports calling through AI tools such as Claude, and can obtain webpage content without API keys.
    2
    2
    Apache 2.0

  • ORZ MCP

    sunwu51
    F
    license
    A
    quality
    D
    maintenance
    Provides web search capabilities across multiple engines and webpage content fetching to simplified Markdown. It enables AI assistants to access real-time internet information and extract text from specific URLs.
    2
    13
    -