byrdocs-search-mcp
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., "@byrdocs-search-mcpsearch for probability theory past exam papers"
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.
byrdocs-search-mcp
北邮 byrdocs 资料 / 真题检索的 serverless MCP,跑在 Cloudflare Workers 上。 任何支持 MCP 的 AI 客户端(Claude Desktop、Cursor、各种 agent)插上就能搜北邮资料和历年真题。
检索逻辑移植自 superdocs-agent,去掉 Mastra 壳、分词换成 WASM 版 jieba。
暴露的工具
工具 | 作用 |
| 搜 byrdocs 资料元信息(教材/试卷/资料),返回标题、课程、学年(试卷为学年区间如 |
| 一步搜新生答疑知识库(生存指南 + 真题wiki),返回可直接据此作答的正文 + 出处 url( |
Related MCP server: Search Docs MCP
架构(为什么这么搭)
MCP 客户端 ──/mcp(Streamable HTTP)──▶ Cloudflare Worker(McpAgent)
│ 冷启动:jieba-wasm 初始化 + minisearch 现建索引
▼
打包进 Worker 的 data/metadata.json + knowledge.json无服务器:没请求不占资源、不烧钱;有请求平台临时拉起。
数据打包进 Worker:
data/*.json(~1.5MB)+ jieba wasm(~4MB),gzip 后约 ~2MB,在免费版 3MB 上限内。数据再涨就改放 R2 运行时拉。分词用
jieba-wasm:Worker 跑不了原生nodejs-jieba,但 WASM 能跑(byrdocs 前端也用它)。
本地跑
npm install
npm run dev # wrangler dev,起本地 Worker
# 访问 http://localhost:8787/ 看说明;MCP 端点是 /mcp 和 /sse部署
npx wrangler login # 登录 Cloudflare 账号(免费即可)
npm run deploy # 部署,拿到 https://byrdocs-search-mcp.<你的子域>.workers.dev接到 Claude Desktop
claude_desktop_config.json 里加(用 mcp-remote 桥接 HTTP MCP):
{
"mcpServers": {
"byrdocs-search": {
"command": "npx",
"args": ["mcp-remote", "https://byrdocs-search-mcp.<你的子域>.workers.dev/sse"]
}
}
}刷新数据
真题/指南更新或资料库变动后:
npm run build-data # 重拉 metadata + 从 GitHub 现建 knowledge(需能访问 GitHub)
npm run deploy # 把新数据打进 Worker也可以挂 GitHub Action 定期跑
build-data+deploy,让数据自动保鲜。
已本地验证 ✅
wrangler dev 本地端到端跑通:listTools 返回两个工具,search_documents/answer_knowledge 真调都返回正确结果(试卷学年区间+阶段、知识库线上 wiki 链接都对)。以下都已解决:
jieba-wasm:包的
exports挡了直接 subpath 导入,故由scripts/copy-wasm.mjs(postinstall/predev/predeploy 自动跑)把pkg/web/jieba_rs_wasm_bg.wasm拷进src/再本地导入。src/*.wasm已 gitignore(装依赖时自动生成)。McpAgent Durable Object:
wrangler.jsonc的MCP_OBJECT绑定 +new_sqlite_classesmigration 本地正常起。依赖版本:
package.json的版本已npm install验证可解析。
唯一没实测的是线上部署本身(需要你的 Cloudflare 账号):wrangler deploy 会报 gzip 后大小,若超免费版 3MB,把 data/ + wasm 改放 R2 运行时拉(见架构说明)。
许可
检索逻辑源自 superdocs-agent;真题/指南内容版权归 byrdocs 及各源仓库(CC-BY-NC-SA 等),本项目只做检索索引,不再分发原文全文。
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
- 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/Jinchen-Yang/byrdocs-search-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server