doubao-search-mcp
This server provides web search capabilities for AI agents via Doubao Search (豆包搜索), with strong Chinese content coverage and rich, long-form results.
Web Search: Submit natural language queries (Chinese or English) and retrieve up to 20 web results with rich metadata.
Exclusive ByteDance Sources: Access Toutiao (今日头条) articles and Douyin Baike (抖音百科), ideal for Chinese entity/concept lookups.
Cross-Language Search: English queries surface first-party sources like official documentation and changelogs.
Long-Form Snippets: Each result includes up to 2000 characters of article body text — ready for direct AI consumption without a separate crawl step.
Fresh, Timestamped Results: Precise publish timestamps (down to the second) and traceable source URLs to assess information freshness.
Configurable Result Count: Return between 1–20 results per query (default: 10).
Adjustable Snippet Length: Control text returned per result (50–2000 characters, default: 600) to manage context window usage.
Token Count per Result: Each result includes a
ContentTokenCountfield for precise context window budgeting.Optional Images: Retrieve up to 3 CDN-hosted image URLs per result (default: 0).
Free Tier: 500 free searches per month, pay-as-you-go beyond that.
Ideal use cases: Fact-checking, recent news, Chinese internet research, entity lookups, and general web research.
Integrates with ByteDance's Doubao Search API to provide web search capabilities, including long-form snippets, publish timestamps, and access to exclusive ByteDance sources like Toutiao and Douyin Baike.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@doubao-search-mcpsearch for latest Chinese AI news"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
doubao-search-mcp
「你给 Claude Code 换上了国产模型,然后发现它不会上网了」
一条命令,让你的 Agent 用上豆包搜索:字节系独家信源、千字级正文摘要、发布时间精确到秒,每月 500 次免费。
豆包搜索是火山引擎为 AI Agent 构建的联网信息服务,2026 年 7 月起面向企业和开发者开放。
为什么需要它
Claude Code 内置的 WebSearch 是 Anthropic 的服务端工具,跑在 Anthropic 的服务器上。一旦你把 ANTHROPIC_BASE_URL 指向国产模型的兼容端点(豆包 Seed、Kimi、GLM、DeepSeek……),WebSearch 就没了执行器:工具看着还在,agent 实际上断网了。很多人是切完模型才发现这件事。
海外搜索 API 能补上这个缺口,但对中文开发者有三个现实问题:美元结算(有的还强制绑信用卡)、Google 系索引搜不到中文生态内容、国内直连不稳定。
豆包搜索正好把这三个问题一起解决了,这个 MCP server 把它变成任何 agent 一条命令就能装上的搜索工具。
Related MCP server: websearch-deepseek
效果
以下是真实返回(工具调用一次的原始输出节选):
共 20 条结果,返回前 10 条:
[1] "一人公司"与"手搓经济",两个新词寓春意-新华网
来源: 新华网 | 2026-03-19T09:00:22+08:00
URL: http://www.xinhuanet.com/tech/20260319/259b0a6eb28c...
当AI技术高歌猛进,"机器换人"的焦虑如影随形。近期走红的"一人公司"与
"手搓经济"两个新词,则为我们揭示了技术进步的另一面……(千字级正文摘要)几个实测过的细节:
千字级正文摘要,不是两行摘要加一堆蓝链接。别家要「搜索 API + 抓取 API」两跳才能凑齐的链路,它一次返回
发布时间精确到秒,agent 自己就能判断信息新鲜度。实测查行业热点,返回过发布时间是「查询前一晚」的文章
字节系独家信源:今日头条正文高频返回;查「XX 是什么」「XX 是谁」这类实体问题,抖音百科基本排第一
每条结果自带正文 token 计数(
ContentTokenCount),你可以精确控制塞进上下文的量——这个字段暴露了它确实是为 Agent 设计的跨语言:英文查询返回官方文档、changelog 这类一手信源,不用中英文各配一个搜索
安装
第一步:拿一个 API Key
去 火山引擎豆包搜索控制台 开通服务并创建 API Key。每月 500 次搜索免费,超出按量付费。
第二步:加进你的 Agent
Claude Code(一条命令):
claude mcp add doubao-search \
-e DOUBAO_SEARCH_API_KEY=你的Key \
-- npx -y github:alchaincyf/doubao-search-mcp加 --scope user 则所有项目可用。
Cursor / Codex / 其他 MCP client(配置文件):
{
"mcpServers": {
"doubao-search": {
"command": "npx",
"args": ["-y", "github:alchaincyf/doubao-search-mcp"],
"env": {
"DOUBAO_SEARCH_API_KEY": "你的Key"
}
}
}
}最省事的方式:把这个仓库链接直接丢给你的 agent,说一句:
帮我安装这个 MCP:https://github.com/alchaincyf/doubao-search-mcp从源码安装
git clone https://github.com/alchaincyf/doubao-search-mcp.git
cd doubao-search-mcp
npm install && npm run build
claude mcp add doubao-search \
-e DOUBAO_SEARCH_API_KEY=你的Key \
-- node /绝对路径/doubao-search-mcp/dist/index.js工具参数
装好后 agent 多出一个 doubao_search 工具:
参数 | 类型 | 默认 | 说明 |
| string | — | 搜索词,中英文都行,自然语言效果就很好 |
| int 1-20 | 10 | 返回条数 |
| int 50-2000 | 600 | 每条摘要的最大字符数,深度阅读时调大 |
| int 0-3 | 0 | 每条结果最多返回几张图(CDN 直链,带宽高) |
背后的故事
我天天泡在 Claude Code 里干活,也经常把国产模型接进来当主力。撞到搜索断网这堵墙之后查了一圈,现有方案对中文开发者各有各的别扭:美元结算、中文内容搜不到、国内连接不稳。
正好豆包搜索在 7 月对开发者开放了。我开通后先自己测了一轮:返回结构里那个 token 计数字段让我确认这个 API 值得包一层 MCP——它不是把「给人看的搜索结果」转给 agent,是从设计上就在替 agent 的上下文预算着想。于是有了这个仓库。
关于作者
花叔 Huashu — AI Native Coder,独立开发者,代表作:小猫补光灯(AppStore 付费榜 Top1)
平台 | 链接 |
🌐 官网 | |
𝕏 Twitter | |
📺 B站 | |
▶️ YouTube | |
📕 小红书 | |
💬 公众号 | 微信搜「花叔」或扫码关注 ↓ |
许可证
MIT — 随便用,随便改,随便造。

女娲.skill — 蒸馏任何人的思维方式 · 达尔文.skill — 让 Skill 无限进化 · huashu-design — HTML 原生设计 skill
English
"You switched Claude Code to a Chinese LLM — and it can't browse the web anymore."
Claude Code's built-in WebSearch is an Anthropic server-side tool. Point ANTHROPIC_BASE_URL at any Anthropic-compatible endpoint (Doubao Seed, Kimi, GLM, DeepSeek...) and your agent silently goes offline.
doubao-search-mcp plugs that gap with Doubao Search — the search API Volcengine built for AI agents, now open to developers with 500 free searches/month:
Long-form snippets (up to 2000 chars of article body per result) — consumable directly, no second crawl needed
Publish time down to the second + source name on every result
Exclusive ByteDance sources: Toutiao articles, Douyin Baike (top-ranked for entity lookups)
Cross-language: English queries return first-party sources (official docs, changelogs)
ContentTokenCountper result — budget your context window precisely
Install (Claude Code):
claude mcp add doubao-search \
-e DOUBAO_SEARCH_API_KEY=your-key \
-- npx -y github:alchaincyf/doubao-search-mcpOr add the JSON config above to any MCP client. Tool: doubao_search(query, count, snippet_length, images).
MIT License © 花叔 Huashu
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Tools
Latest Blog Posts
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/alchaincyf/doubao-search-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server