docpilot-mcp

docpilot-mcp

An MCP server that provides tools to fetch live, version-accurate documentation, changelogs, examples, and method signatures for npm and PyPI packages, preventing AI coding agents from hallucinating stale APIs.

Category
访问服务器

README

docpilot-mcp

A TypeScript Model Context Protocol (MCP) server that exposes 4 tools so any AI coding agent (Claude Code, Cursor, Codex, etc.) can fetch live, version-accurate documentation and code examples for any npm or PyPI package — and never again hallucinate a stale method signature.

Tool What it does
query_docs Answer a natural-language question with a cited, ranked context block from the official docs
get_changelog Return the 10 most recent changelog entries for a package
search_examples Return up to 10 real code examples from the package's official examples/ dir or README.md
resolve_method Return the current signature + parameter list for a method in a pinned version

The server runs locally on stdio, persists a vector index in ~/.cache/docpilot-mcp/, and never calls anything other than the package's own registry, GitHub repo, and docs site.


Requirements

  • Node.js ≥ 20 (this project is tested on Node 24)
  • npm ≥ 11 (any package manager that understands the package.json exports map)
  • For JS-rendered doc sites (Stripe, Vercel, Cloudflare, etc.) the first call will need a one-time npx playwright install chromium download (~150 MB)

Install

git clone <this repo> docpilot-mcp
cd docpilot-mcp
npm install
npm run build

npm install is fully deterministic against the locked package-lock.json. npm run build produces dist/ and exits 0.

Run

# Foreground stdio server (consumed by an MCP client)
npm start

# Dev mode with tsx hot reload
npm run dev

# Enumerate the 4 tools + their descriptions
npm run inspect

On startup the server writes a single line to stderr:

info docpilot-mcp ready (tools=4)

It then waits for JSON-RPC messages on stdin. When the parent MCP client closes stdin the server exits with code 0.

Pointing a client at the server

Once the package is published:

# Claude Code
claude mcp add docpilot -- npx -y docpilot-mcp

# Cursor / Codex (config file)
#   command:    npx
#   args:       -y
#   other args: docpilot-mcp

From a local checkout before publishing:

# Claude Code
claude mcp add docpilot -- node /abs/path/to/docpilot-mcp/dist/index.js

Tool examples

query_docs

{
  "package": "stripe",
  "version": "5.0.0",
  "question": "how do I paginate cursor results"
}

Returns a markdown answer (≤ 2000 tokens) plus a sources: [{ url, section, snippet, score }] array. The top snippet is always a literal phrase from the package's own docs.

get_changelog

{ "package": "stripe", "version": "latest" }

Returns the 10 most recent entries from the npm registry metadata, with a transparent fallback to CHANGELOG.md on the default branch.

search_examples

{
  "package": "stripe",
  "version": "5.0.0",
  "query": "create a customer"
}

Returns up to 10 { code, path, url, language } blocks. Every url is on github.com under the package's official repo.

resolve_method

{
  "package": "stripe",
  "version": "5.0.0",
  "method": "customers.create"
}

Returns { signature, params, returns, source: { url, path, line } }. The signature is parsed from the .d.ts file inside the npm tarball; the line is 1-indexed.

Caching

  • Default cache root: ~/.cache/docpilot-mcp/
  • Layout: index/<ecosystem>/<package>/<version>/ for the vector store, plus a raw/ tree for cached HTML and tarballs
  • All disk writes are atomic (*.tmp → rename)
  • See docs/cache-format.md (added in AC-7) for the full schema

Testing

npm test            # runs vitest run
npm run test:watch  # watch mode

The test suite uses recorded fixtures under test/fixtures/ so it stays fully offline and deterministic.

Project layout

src/
  index.ts              # entry point — wires transport + handlers
  server.ts             # MCP Server factory + tool registration table
  tools/                # the 4 tool handlers (one file per tool)
  sources/              # fetchPage, registry clients, GitHub resolver
  extractors/           # TS / Python / markdown chunkers
  index/                # embeddings + hnswlib store
  browser/              # playwright singleton pool
  net/                  # rate-limited, retrying HTTP client
  cache/                # content-addressed paths
  util/                 # logger, error helpers
test/                   # vitest specs + fixtures
docs/
  references/           # recorded raw materials (browser snapshots, registry JSON)
  cache-format.md
  verified-packages.md

License

MIT

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选