Github的维基百科DeepWiKi
📖 MCP 服务器用于获取 deepwiki.com 的最新知识，并在 Cursor 和其他代码编辑器中使用
MCP 服务配置
复制以下 JSON 到 OPClaw 或其他 MCP 客户端的配置文件中即可使用
{
"mcpServers": {
"mcp-deepwiki": {
"args": [
"-y",
"mcp-deepwiki@latest"
],
"command": "npx"
}
}
}可用工具 （1 个）
该服务在 MCP 协议中暴露的工具，AI 可按需调用
deepwiki_fetch 4 个参数 需填 1 项
Fetch a deepwiki.org repo and return Markdown
必填参数：url
服务介绍
Deepwiki MCP Server
This is an unofficial Deepwiki MCP server.
It receives a Deepwiki URL via MCP, scrapes all related pages, converts them to Markdown format, and returns a document or a list of pages.
Features
- 🔒 Domain Security: Only handles URLs from deepwiki.org
- 🧹 HTML Cleanup: Removes headers, footers, navigation, scripts, and ads
- 🔗 Link Rewriting: Adjusts links to work in Markdown
- 📄 Multiple Output Formats: Get a single document or structured pages
- 🚀 Performance: Fast scraping with adjustable concurrency and depth
- NLP: Used only for library name search
Usage
Prompts you can use:
deepwiki fetch how can i use gpt-image-1 with "vercel ai" sdk
deepwiki fetch how can i create new blocks in shadcn?
deepwiki fetch i want to understand how X works
Get the full document (default)
use deepwiki https://deepwiki.org/shadcn-ui/ui
use deepwiki multiple pages https://deepwiki.org/shadcn-ui/ui
Single page
use deepwiki fetch single page https://deepwiki.org/tailwindlabs/tailwindcss/2.2-theme-system
Use short form to get
use deepwiki fetch tailwindlabs/tailwindcss
deepwiki fetch library
deepwiki fetch url
deepwiki fetch <name>/<repo>
deepwiki multiple pages ...
deepwiki single page url ...
Cursor
Add the following to your .cursor/mcp.json file.
{
"mcpServers": {
"mcp-deepwiki": {
"command": "npx",
"args": [
"-y",
"mcp-deepwiki@latest"
]
}
}
}
MCP Tool Integration
This package registers a tool named deepwiki_fetch that you can use in any MCP-compatible client:
{
"action": "deepwiki_fetch",
"params": {
"url": "https://deepwiki.org/user/repo",
"mode": "aggregate",
"maxDepth": "1"
}
}
Parameters
url(required): The starting URL of the Deepwiki repositorymode(optional): Output mode, can be "aggregate" for a single Markdown document (default), or "pages" for structured page datamaxDepth(optional): Maximum page depth to scrape (default: 10)
Response Format
Successful Response (Aggregate Mode)
{
"status": "ok",
"data": "# 页面标题
页面内容...
---
# 另一个页面
更多内容...",
"totalPages": 5,
"totalBytes": 25000,
"elapsedMs": 1200
}
Successful Response (Pages Mode)
{
"status": "ok",
"data": [
{
"path": "index",
"markdown": "# 主页
欢迎来到仓库。"
},
{
"path": "section/page1",
"markdown": "# 第一页
这是第一页的内容。"
}
],
"totalPages": 2,
"totalBytes": 12000,
"elapsedMs": 800
}
Error Response
{
"status": "error",
"code": "DOMAIN_NOT_ALLOWED",
"message": "仅允许 deepwiki.org 域名"
}
Partial Success Response
{
"status": "partial",
"data": "# 页面标题
页面内容...",
"errors": [
{
"url": "https://deepwiki.org/user/repo/page2",
"reason": "HTTP 错误: 404"
}
],
"totalPages": 1,
"totalBytes": 5000,
"elapsedMs": 950
}
Progress Events
While using the tool, you will receive progress events during the scraping process:
Fetched https://deepwiki.org/user/repo: 12500 bytes in 450ms (status: 200)
Fetched https://deepwiki.org/user/repo/page1: 8750 bytes in 320ms (status: 200)
Fetched https://deepwiki.org/user/repo/page2: 6200 bytes in 280ms (status: 200)
Local Development - Installation
Local Use
{
"mcpServers": {
"mcp-deepwiki": {
"command": "node",
"args": [
"./bin/cli.mjs"
]
}
}
}
Install from Source
# 克隆仓库
git clone https://github.com/regenrek/mcp-deepwiki.git
cd mcp-deepwiki
# 安装依赖
npm install
# 构建包
npm run build
Direct API Call
For HTTP transport, you can make a direct API call:
curl -X POST http://localhost:3000/mcp
-H "Content-Type: application/json"
-d '{
"id": "req-1",
"action": "deepwiki_fetch",
"params": {
"url": "https://deepwiki.org/user/repo",
"mode": "aggregate"
}
}'
Configuration
Environment Variables
DEEPWIKI_MAX_CONCURRENCY: Maximum number of concurrent requests (default: 5)DEEPWIKI_REQUEST_TIMEOUT: Request timeout in milliseconds (default: 30000)DEEPWIKI_MAX_RETRIES: Maximum number of retries for failed requests (default: 3)DEEPWIKI_RETRY_DELAY: Base delay for retry backoff in milliseconds (default: 250)
To configure these environment variables, create a .env file in the root of the project:
DEEPWIKI_MAX_CONCURRENCY=10
DEEPWIKI_REQUEST_TIMEOUT=60000
DEEPWIKI_MAX_RETRIES=5
DEEPWIKI_RETRY_DELAY=500
Docker Deployment (Untested)
Build and run the Docker image:
# 构建镜像
docker build -t mcp-deepwiki .
# 使用 stdio 传输方式运行（适用于开发）
docker run -it --rm mcp-deepwiki
# 使用 HTTP 传输方式运行（适用于生产）
docker run -d -p 3000:3000 mcp-deepwiki --http --port 3000
# 使用环境变量运行
docker run -d -p 3000:3000
-e DEEPWIKI_MAX_CONCURRENCY=10
-e DEEPWIKI_REQUEST_TIMEOUT=60000
mcp-deepwiki --http --port 3000
Development
# 安装依赖
pnpm install
# 使用 stdio 方式以开发模式运行
pnpm run dev-stdio
# 运行测试
pnpm test
# 运行代码检查
pnpm run lint
# 构建包
pnpm run build
Troubleshooting
Common Issues
-
Permission Denied: If you encounter an EACCES error when running the CLI, ensure the binary is executable:
bash
chmod +x ./node_modules/.bin/mcp-deepwiki -
Connection Refused: Ensure the port is available and not blocked by a firewall:
bashCheck if the port is in use
lsof -i :3000
-
Timeout Errors: For large repositories, consider increasing the timeout and concurrency:
bash
DEEPWIKI_REQUEST_TIMEOUT=60000 DEEPWIKI_MAX_CONCURRENCY=10 npx mcp-deepwiki
Contributing
We welcome contributions! See CONTRIBUTING.md for details.
License
MIT
Links
- X/Twitter: @kregenrek
- Bluesky: @kevinkern.dev
Courses
- Learn Cursor AI: Ultimate Cursor Course
- Learn to Build Software with AI: instructa.ai
Check Out My Other Projects:
- AI Prompts - Curated AI prompts for Cursor AI, Cline, Windsurf, and Github Copilot
- codefetch - Convert code to Markdown format with a simple terminal command, making it easy for LLMs to use* aidex - A CLI tool that provides detailed information about AI language models, helping developers choose the right model for their needs.
