基于 DeepSeek 网页版的在线搜索 MCP 服务器。通过 DeepSeek 的搜索增强对话功能，为 MCP 客户端提供实时网络搜索能力。
提供 web_search MCP 工具，支持实时网络搜索
支持 DeepSeek 默认模型和专家模型
支持思考过程（thinking）输出
基于 MCP SDK，stdio 传输，兼容所有 MCP 客户端
stdio 传输，兼容所有 MCP 客户端
Related MCP server: google-search-mcp
cd web-search-mcp
npm install编辑 Claude Desktop 配置文件（claude_desktop_config.json），通过 env 字段直接传入凭据，无需 .env 文件：
{
"mcpServers": {
"web-search": {
"command": "node",
"args": ["path/to/web-search-mcp/src/index.js"],
"env": {
"DEEPSEEK_AUTH_TOKEN": "your_token_here"
}
}
}
}在项目的 .claude/settings.json 中添加：
{
"mcpServers": {
"web-search": {
"command": "node",
"args": ["path/to/web-search-mcp/src/index.js"],
"env": {
"DEEPSEEK_AUTH_TOKEN": "your_token_here"
}
}
}
}如果不方便在 MCP 配置中写 env，也可以用 .env 文件：
cp .env.example .env
# 编辑 .env 填入凭据环境变量优先级：系统环境变量 > MCP
env配置 >.env文件
web_search使用 DeepSeek 搜索增强对话进行网络搜索，返回基于搜索结果的回答。
参数：
参数 | 类型 | 必填 | 说明 |
| string | 是 | 搜索查询内容 |
| boolean | 否 | 是否开启 DeepSeek 思考/推理模式，开启后模型会先推理再回答，推理过程包含在响应中（默认 false） |
| string | 否 | 模型模式： |
变量 | 必填 | 默认值 | 说明 |
| 二选一 | - | DeepSeek 登录 token |
| 二选一 | - | DeepSeek 浏览器 cookie |
| 否 |
| 模型类型： |
| 否 |
| 是否默认启用思考过程 |
| 否 |
| PoW 求解器 WASM 文件路径 |
| 否 |
| 日志级别： |
打开 chat.deepseek.com 并登录
打开浏览器开发者工具（F12）→ 切换到 Console（控制台） 面板
输入以下命令并回车，即可复制 token：
copy(JSON.parse(localStorage.getItem("userToken")).value)粘贴到 DEEPSEEK_AUTH_TOKEN 环境变量中即可
注意：凭据会过期，过期后需要重新获取。Token 过期的典型表现是请求返回 Cloudflare 验证页面。
# 运行所有测试
npm test
# 运行单个测试文件
node --test tests/mcp-protocol.test.js
node --test tests/config.test.jssrc/
index.js # 入口，初始化所有组件并启动 MCP 服务器
mcp-server.js # MCP 服务器（基于 @modelcontextprotocol/sdk，注册 web_search 工具）
deepseek-client.js # DeepSeek 网页版 API 客户端（会话管理、PoW、SSE 流式解析）
pow-solver.js # SHA3 PoW WASM 求解器
env.js # .env 文件加载器
config.js # 环境变量配置解析Search the web using DeepSeek's search-enabled chat. Returns a search-grounded response with up-to-date information.
| Name | Required | Description | Default |
|---|---|---|---|
| query | Yes | The search query | |
| thinking | No | Enable DeepSeek thinking/reasoning mode. When true, the model will reason through the query before answering, and the reasoning trace is included in the response. (default: false) | |
| model | No | DeepSeek model mode (default: from server config) |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The description states that the tool returns a 'search-grounded response with up-to-date information', which implies read-only behavior, but it does not explicitly disclose side effects, authentication needs, or limitations. Without annotations, the description carries a heavier burden and falls short of full transparency.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single sentence that conveys the essential function without unnecessary words. While it is concise, it could be slightly expanded to improve clarity without becoming verbose.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given that there is no output schema, the description should elaborate on the return format. It mentions 'search-grounded response' but does not specify structure or content. The lack of sibling tools reduces the need for differentiation, but completeness is only moderate.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has 100% description coverage for all three parameters, so the baseline is 3. The description adds no extra meaning beyond what the schema already provides for the parameters.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description uses a specific verb 'Search' and clearly identifies the resource 'the web using DeepSeek's search-enabled chat', leaving no ambiguity about what the tool does.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No guidance is provided on when to use this tool versus any alternatives, nor are there any conditions or prerequisites mentioned. The description lacks context for appropriate usage.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v0.1.0web_searchScored across 1 tool
Only one tool exists, so there is no possibility of confusion between tools.
The single tool name 'web_search' is clear and follows a common verb_noun pattern, but with only one tool, consistency cannot be fully assessed.
One tool is minimal for a web search server. It covers the core action but may lack supporting tools for refined queries or result handling.
The tool provides basic web search functionality, but there are no tools for advanced filtering, pagination, or retrieving structured results, leaving potential gaps.
Live AI-native web search with citations. One tool for every MCP client. Flat per-request pricing.
Web search, scraping, RAG answers with citations, and translation as MCP tools.
Free web search for AI agents. No API key required. Hosted MCP in active development.
Search the agentic web. 4,100+ sites, 11 tools incl. check_url + verify_mcp for probe-before-use.
