MCP Rust Docs Server

MCP Rust Docs Server

A Model Context Protocol (MCP) server for fetching Rust crate documentation from docs.rs using the rustdoc JSON API.

Category
访问服务器

README

🦀 MCP Rust Docs Server

MCP Protocol Rust Docs TypeScript Bun License: Apache 2.0

A Model Context Protocol (MCP) server for fetching Rust crate documentation from docs.rs using the rustdoc JSON API

FeaturesInstallationUsageBuildingDevelopmentNotesContributingLicense

✨ Features

<a id="features"></a>

  • 🚀 Fast Documentation Fetching - Direct access to rustdoc JSON API for comprehensive crate documentation
  • 🔍 Item-Level Lookup - Query specific structs, functions, traits, and more within crates
  • 💾 Smart Caching - Built-in LRU cache with SQLite backend for optimal performance
  • 🎯 Version Support - Fetch docs for specific versions or use semver ranges
  • 🖥️ Cross-Platform - Standalone executables for Linux, macOS, and Windows
  • 📦 Zero Dependencies - Single executable with everything bundled
  • 🔧 TypeScript - Full type safety with modern ES modules
  • 🗜️ Compression Support - Automatic Zstd decompression for efficient data transfer

📦 Installation

<a id="installation"></a>

Using Bun

bun install
bun run build:bytecode # or bun run build:all for all platforms

Using Pre-built Executables

Download the latest release for your platform from the Releases page:

Linux

  • x64/AMD64 (GLIBC): mcp-docsrs-linux-x64 - For Ubuntu, Debian, Fedora, etc.
  • ARM64 (GLIBC): mcp-docsrs-linux-arm64 - For ARM64 systems, AWS Graviton
  • x64/AMD64 (MUSL): mcp-docsrs-linux-x64-musl - For Alpine Linux, Docker containers (requires libstdc++)
  • ARM64 (MUSL): mcp-docsrs-linux-arm64-musl - For Alpine on ARM64, minimal containers (requires libstdc++)

macOS

  • Intel: mcp-docsrs-darwin-x64 - For Intel-based Macs
  • Apple Silicon: mcp-docsrs-darwin-arm64 - For M1/M2/M3 Macs

Windows

  • x64: mcp-docsrs-windows-x64.exe - For 64-bit Windows

Using Docker

Pull and run the latest multi-arch image (supports both x64 and ARM64):

# Pull the latest image
docker pull ghcr.io/vexxvakan/mcp-docsrs:latest

# Run the server
docker run --rm -i ghcr.io/vexxvakan/mcp-docsrs:latest

# Run with custom configuration
docker run --rm -i ghcr.io/vexxvakan/mcp-docsrs:latest \
  --cache-ttl 7200000 --max-cache-size 200

Available tags:

  • latest - Latest stable release (multi-arch)
  • v1.0.0 - Specific version (multi-arch)
  • x64 - Latest x64/AMD64 build
  • arm64 - Latest ARM64 build

🚀 Usage

<a id="usage"></a>

Starting the Server

Using npm or Bun

# Production mode
npm start
# or
bun start

# Development mode with hot reload
npm run dev
# or
bun run dev

Using Executable

# Show help
mcp-docsrs --help

# Run with default settings
mcp-docsrs

# Run with custom configuration
mcp-docsrs --cache-ttl 7200000 --max-cache-size 200

🛠️ Available Tools

lookup_crate_docs

Fetches comprehensive documentation for an entire Rust crate.

Parameters:

Parameter Type Required Description
crateName string Name of the Rust crate
version string Specific version or semver range (e.g., "1.0.0", "~4")
target string Target platform (e.g., "i686-pc-windows-msvc")
formatVersion string Rustdoc JSON format version

Example:

{
  "tool": "lookup_crate_docs",
  "arguments": {
    "crateName": "serde",
    "version": "latest"
  }
}

lookup_item_docs

Fetches documentation for a specific item within a crate.

Parameters:

Parameter Type Required Description
crateName string Name of the Rust crate
itemPath string Path to the item (e.g., "struct.MyStruct", "fn.my_function")
version string Specific version or semver range
target string Target platform

Example:

{
  "tool": "lookup_item_docs",
  "arguments": {
    "crateName": "tokio",
    "itemPath": "runtime.Runtime"
  }
}

📊 Resources

The server provides resources for querying and inspecting the cache database:

cache://stats

Returns cache statistics including total entries, size, and oldest entry.

Example:

{
  "totalEntries": 42,
  "totalSize": 1048576,
  "oldestEntry": "2024-01-15T10:30:00.000Z"
}

cache://entries?limit={limit}&offset={offset}

Lists cached entries with metadata. Supports pagination.

Parameters:

  • limit - Number of entries to return (default: 100)
  • offset - Number of entries to skip (default: 0)

Example:

[
  {
    "key": "serde/latest/x86_64-unknown-linux-gnu",
    "timestamp": "2024-01-15T14:20:00.000Z",
    "ttl": 3600000,
    "expiresAt": "2024-01-15T15:20:00.000Z",
    "size": 524288
  }
]

cache://query?sql={sql}

Execute SQL queries on the cache database (SELECT queries only for safety).

Example:

cache://query?sql=SELECT key, timestamp FROM cache WHERE key LIKE '%tokio%' ORDER BY timestamp DESC

Note: SQL queries in the URI should be URL-encoded. The server will automatically decode them.

cache://config

Returns the current server configuration including all runtime parameters.

Example response:

{
  "cacheTtl": 7200000,
  "maxCacheSize": 200,
  "requestTimeout": 30000,
  "dbPath": "/Users/vexx/Repos/mcp-docsrs/.cache"
}

⚙️ Configuration

Configure the server using environment variables or command-line arguments:

Variable CLI Flag Default Description
CACHE_TTL --cache-ttl 3600000 Cache time-to-live in milliseconds
MAX_CACHE_SIZE --max-cache-size 100 Maximum number of cached entries
REQUEST_TIMEOUT --request-timeout 30000 HTTP request timeout in milliseconds
DB_PATH --db-path :memory: Path to SQLite database file (use :memory: for in-memory)

Example:

# Environment variables
CACHE_TTL=7200000 MAX_CACHE_SIZE=200 npm start

# Command-line arguments (executable)
./mcp-docsrs --cache-ttl 7200000 --max-cache-size 200

# Use persistent database to cache documentation between sessions
./mcp-docsrs --db-path ~/.mcp-docsrs

# Or with environment variable
DB_PATH=~/.mcp-docsrs npm start

🔌 MCP Configuration

Add to your MCP configuration file:

{
  "mcpServers": {
    "rust-docs": {
      "command": "node",
      "args": ["/path/to/mcp-docsrs/dist/index.js"]
    }
  }
}

Or using the executable:

{
  "mcpServers": {
    "rust-docs": {
      "command": "/path/to/mcp-docsrs"
    }
  }
}

Or using Docker:

{
  "mcpServers": {
    "rust-docs": {
      "command": "docker",
      "args": ["run", "--rm", "-i", "ghcr.io/vexxvakan/mcp-docsrs:latest"]
    }
  }
}

<a id="building"></a>

🏗️ Building

<a id="building"></a>

Prerequisites

  • Bun v1.2.14 or later
  • macOS, Linux, or Windows

Build Commands

# Build for current platform
bun run build

# Build with bytecode compilation (standalone, requires Bun runtime)
bun run build:bytecode

# Build for all platforms (7 targets, all with bytecode for fast startup)
bun run build:all

# Linux builds (GLIBC - standard)
bun run build:linux-x64      # Linux x64/AMD64
bun run build:linux-arm64    # Linux ARM64

# Linux builds (MUSL - for Alpine/containers)
bun run build:linux-x64-musl    # Linux x64/AMD64 (Alpine)
bun run build:linux-arm64-musl  # Linux ARM64 (Alpine)

# macOS builds
bun run build:darwin-x64     # macOS Intel
bun run build:darwin-arm64   # macOS Apple Silicon

# Windows build
bun run build:windows-x64    # Windows x64

Build Output

All executables are created in the dist/ directory with bytecode compilation for fast startup:

File Platform Type Size
mcp-docsrs-linux-x64 Linux x64/AMD64 GLIBC + Bytecode ~99MB
mcp-docsrs-linux-arm64 Linux ARM64 GLIBC + Bytecode ~93MB
mcp-docsrs-linux-x64-musl Linux x64/AMD64 MUSL (static) + Bytecode ~92MB
mcp-docsrs-linux-arm64-musl Linux ARM64 MUSL (static) + Bytecode ~88MB
mcp-docsrs-darwin-x64 macOS Intel Bytecode ~64MB
mcp-docsrs-darwin-arm64 macOS Apple Silicon Bytecode ~58MB
mcp-docsrs-windows-x64.exe Windows x64 Bytecode ~113MB

<a id="development"></a>

👨‍💻 Development

<a id="development"></a>

Development Workflow

# Install dependencies
bun install

# Run in development mode
bun run dev

# Run tests
bun test

# Lint code
bun run lint

# Type checking
bun run typecheck

# Check build sizes (updates README table)
bun run check:sizes  # Run after building

Testing

The project includes comprehensive tests for all major components:

# Run all tests
bun test

# Run tests in watch mode
bun test --watch

# Run specific test file
bun test cache.test.ts

# Run tests with full error logging (including expected errors)
LOG_EXPECTED_ERRORS=true bun test

Test Output

Tests are configured to provide clean output by default:

  • ✅ Expected errors (like CrateNotFoundError in 404 tests) show as green checkmarks: ✓ Expected CrateNotFoundError thrown
  • ❌ Unexpected errors are shown with full stack traces in red
  • ℹ️ Info logs are shown to track test execution

This makes it easy to distinguish between:

  • Tests that verify error handling (expected errors)
  • Actual test failures (unexpected errors)

To see full error details for debugging, set LOG_EXPECTED_ERRORS=true.

Project Structure

mcp-docsrs/
├── src/                        # Source code
│   ├── cli.ts                  # CLI entry point with argument parsing
│   ├── index.ts                # MCP server entry point
│   ├── server.ts               # MCP server implementation with tool/resource handlers
│   ├── cache.ts                # LRU cache with SQLite persistence
│   ├── docs-fetcher.ts         # HTTP client for docs.rs JSON API
│   ├── rustdoc-parser.ts       # Parser for rustdoc JSON format
│   ├── errors.ts               # Custom error types and error handling
│   ├── types.ts                # TypeScript types and Zod schemas
│   └── tools/                  # MCP tool implementations
│       ├── index.ts            # Tool exports and registration
│       ├── lookup-crate.ts     # Fetch complete crate documentation
│       ├── lookup-item.ts      # Fetch specific item documentation
│       └── search-crates.ts    # Search crates on crates.io
├── test/                       # Test files
│   ├── cache.test.ts           # Cache functionality tests
│   ├── cache-status.test.ts    # Cache status and metrics tests
│   ├── docs-fetcher.test.ts    # API client tests
│   ├── integration.test.ts     # End-to-end integration tests
│   ├── persistent-cache.test.ts # SQLite cache persistence tests
│   ├── rustdoc-parser.test.ts  # JSON parser tests
│   └── search-crates.test.ts   # Crate search tests
├── scripts/                    # Development and testing scripts
│   ├── test-crates-search.ts   # Manual crate search testing
│   ├── test-mcp.ts             # MCP server testing
│   ├── test-persistent-cache.ts # Cache persistence testing
│   ├── test-resources.ts       # Resource endpoint testing
│   └── test-zstd.ts            # Zstandard compression testing
├── plans/                      # Project planning documents
│   └── feature-recommendations.md # Future feature ideas
├── dist/                       # Build output (platform executables)
├── .github/                    # GitHub Actions workflows
│   ├── workflows/              # CI/CD pipeline definitions
│   └── ...                     # Various automation configs
├── CLAUDE.md                   # AI assistant instructions
├── README.md                   # Project documentation
├── LICENSE                     # Apache 2.0 license
├── package.json                # Project dependencies and scripts
├── tsconfig.json               # TypeScript configuration
├── biome.json                  # Code formatter/linter config
└── bun.lock                    # Bun package lock file

<a id="notes"></a>

📝 Notes

<a id="notes"></a>

  • 📅 The rustdoc JSON feature on docs.rs started on 2025-05-23, so releases before that date won't have JSON available
  • 🔄 The server automatically handles redirects and format version compatibility
  • ⚡ Cached responses significantly improve performance for repeated lookups
  • 📦 Built executables include all dependencies - no runtime installation required
  • ⚠️ MUSL builds limitation: Due to a known Bun issue, MUSL builds are not fully static and require libstdc++ to run. For Docker/Alpine deployments, install libstdc++ with: apk add libstdc++

🤝 Contributing

<a id="contributing"></a>

Contributions are welcome! Please feel free to submit a Pull Request.

  1. Fork the repository
  2. Create your feature branch (git checkout -b feature/amazing-feature)
  3. Commit your changes (git commit -m 'Add some amazing feature')
  4. Push to the branch (git push origin feature/amazing-feature)
  5. Open a Pull Request

<a id="license"></a>

📄 License

<a id="license"></a>

This project is licensed under the MIT License - see the LICENSE file for details.


Made with ❤️ for the Rust community

Report BugRequest Feature

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