A Model Context Protocol (MCP) 서버는 AI 코딩 조수 내에서 여러 소스에서 개발자 문서에 원활한 액세스를 제공합니다.

지원 문서 제공자

빠른 시작

Rust 바이너리 빌드 ( Rust 1.76+ 필요):

MCP 클라이언트 구성:

현지 개발:

사용법

Multi-Provider Documentation MCP Server

A Model Context Protocol (MCP) server written in Rust that provides seamless access to developer documentation from multiple sources directly within your AI coding assistant.

Supported Documentation Providers

Quick Start

Build the Rust binary (requires Rust 1.76+):

cargo build --release

Configure your MCP client:

{
  "mcpServers": {
    "docs-mcp": {
      "command": "/absolute/path/to/target/release/docs-mcp-cli"
    }
  }
}

For local development:

cargo run -p docs-mcp-cli

Usage

The server exposes a single unified query tool that automatically detects the appropriate provider and returns comprehensive documentation.

Natural Language Queries

Simply describe what you’re looking for:

query { "query": "SwiftUI NavigationStack" }
query { "query": "Rust tokio spawn async" }
query { "query": "Telegram sendMessage parameters" }
query { "query": "JavaScript Array map filter" }
query { "query": "React useState hook" }
query { "query": "Next.js server components" }
query { "query": "Node.js fs readFile" }

How-To Queries

Ask implementation questions:

query { "query": "how to use SwiftUI NavigationStack" }
query { "query": "how to implement tab navigation in SwiftUI" }

Provider Auto-Detection

The query tool automatically routes to the correct provider based on keywords:

• Apple: SwiftUI, UIKit, iOS, macOS, Foundation, CoreData, etc.
• Rust: std, tokio, serde, HashMap, Vec, async, etc.
• Telegram: bot, sendMessage, getUpdates, webhook, etc.
• TON: blockchain, wallet, jetton, tonapi, etc.
• Cocoon: confidential computing, TDX, attestation, etc.
• MDN: JavaScript, JS, DOM, fetch, promise, array, etc.
• React: hook, useState, useEffect, component, JSX, etc.
• Next.js: nextjs, App Router, server component, etc.
• Node.js: nodejs, fs, path, http, stream, etc.
• MLX: mlx, mlxarray, mlxnn, apple silicon, ml-explore, etc.
• Hugging Face: huggingface, transformers, automodel, autotokenizer, etc.

What You Get

For each query, the tool returns:

• Full documentation content (not truncated summaries)