MkDocs MCP Server
Enables AI assistants to access and search MkDocs documentation through tools for full-text search, page navigation, and code block extraction. It serves documentation pages as readable resources and provides structural outlines to help LLMs navigate documentation content.
README
MkDocs MCP Example
A comprehensive example project demonstrating the integration of MkDocs Material documentation with a Model Context Protocol (MCP) server, showcasing modern Python development practices with uv, devcontainers, and VSCode.
🌟 Features
📚 Beautiful Documentation
- MkDocs Material theme with modern design
- Responsive layout for all devices
- Advanced search with full-text indexing
- Dark/light mode with system preference detection
- Mermaid diagrams and syntax highlighting
- Auto-generated API documentation with mkdocstrings
🤖 AI-Powered Documentation Access
- MCP Server providing AI access to documentation
- Resource serving for direct content access
- Advanced search tools for content discovery
- Code block extraction and analysis
- Page outline generation and navigation
🛠️ Modern Development Stack
- Python 3.11+ with type hints and modern practices
- uv for lightning-fast dependency management
- DevContainers for consistent development environments
- VSCode integration with comprehensive tooling
- Rootless Podman support for secure containerization
🧪 Quality Assurance
- Comprehensive testing with pytest and coverage
- Code formatting with Ruff
- Type checking with MyPy
- Pre-commit hooks for automated quality checks
- CI/CD ready configuration
🚀 Quick Start
Prerequisites
- Podman (rootless preferred)
- VSCode with Dev Containers extension
- Git
1. Clone & Open
git clone https://github.com/robmatesick/mkdocs-mcp-example.git
cd mkdocs-mcp-example
code .
2. Start DevContainer
- Press
Ctrl+Shift+P(Windows/Linux) orCmd+Shift+P(macOS) - Select "Dev Containers: Reopen in Container"
- Wait for container setup (~5-10 minutes on first run)
3. Start Development
# Terminal 1: Documentation server
make docs-serve
# Terminal 2: MCP server
make mcp-server
Visit http://localhost:8000 to see your documentation!
📁 Project Structure
mkdocs-mcp-example/
├── 📁 docs/ # Documentation source
│ ├── index.md # Homepage
│ ├── getting-started/ # Setup guides
│ ├── mcp-server/ # MCP documentation
│ ├── api/ # Auto-generated API docs
│ └── examples/ # Usage examples
│
├── 📁 mkdocs-site/ # MkDocs configuration
│ ├── mkdocs.yml # Main configuration
│ ├── pyproject.toml # Dependencies
│ └── docs/ # Theme customizations
│
├── 📁 mcp-server/ # MCP server implementation
│ ├── src/mkdocs_mcp/ # Source code
│ │ ├── server.py # Main server
│ │ ├── resources.py # Resource management
│ │ └── tools.py # Search tools
│ ├── tests/ # Unit tests
│ └── pyproject.toml # Dependencies
│
├── 📁 .devcontainer/ # DevContainer config
├── 📁 .vscode/ # VSCode settings
├── 📁 tests/ # Integration tests
├── pyproject.toml # Workspace configuration
├── Makefile # Development commands
└── README.md # This file
🔧 Development Commands
The project includes a comprehensive Makefile with common development tasks:
📦 Setup & Installation
make setup # Complete development setup
make install # Install dependencies only
🧹 Code Quality
make format # Format code with Ruff
make lint # Lint code and fix issues
make typecheck # Run MyPy type checking
make quality # Run all quality checks
🧪 Testing
make test # Run all tests
make test-cov # Run tests with coverage
make test-watch # Run tests in watch mode
📚 Documentation
make docs-serve # Start documentation server
make docs-build # Build static documentation
make docs-clean # Clean build artifacts
🤖 MCP Server
make mcp-server # Start MCP server
make mcp-test # Run MCP server tests
🚀 Development Workflow
make dev # Start both docs and MCP server
make clean # Clean all build artifacts
make ci-check # Run all CI checks
🏗️ Architecture
graph TB
A[📚 MkDocs Site] --> B[📄 Documentation Files]
B --> C[🤖 MCP Server]
C --> D[🧠 AI Assistant]
D --> E[👥 Users]
F[👨💻 Developer] --> G[🐳 DevContainer]
G --> H[📦 uv Environment]
H --> A
H --> C
I[🔄 CI/CD] --> J[🧪 Tests]
I --> K[🏗️ Build]
I --> L[🚀 Deploy]
🤖 MCP Server Usage
The MCP server provides AI assistants with direct access to your documentation:
Available Resources
- Documentation pages as readable resources
- Automatic metadata extraction from frontmatter
- Hierarchical navigation support
Available Tools
search_docs- Full-text search across documentationfind_by_title- Find pages by title or headinglist_pages- List all available documentation pagesget_page_outline- Extract page structure and headingssearch_code_blocks- Find and filter code examples
Example Usage
from mcp import ClientSession, StdioServerParameters
from mcp.client.stdio import stdio_client
# Connect to MCP server
server_params = StdioServerParameters(
command="python",
args=["-m", "mkdocs_mcp.server"]
)
async with stdio_client(server_params) as (read, write):
async with ClientSession(read, write) as session:
# List available documentation
resources = await session.list_resources()
print(f"Found {len(resources.resources)} pages")
# Search documentation
result = await session.call_tool("search_docs", {
"query": "installation",
"max_results": 5
})
print(result.content[0].text)
🏃♂️ Getting Started Guide
For Documentation Writers
- Edit content in the
docs/directory - Add new pages and update navigation in
mkdocs.yml - Preview changes at http://localhost:8000
- Use Markdown features like admonitions, code blocks, and diagrams
For Python Developers
- Modify MCP server in
mcp-server/src/mkdocs_mcp/ - Add new tools or resources for AI access
- Run tests with
make test - Follow type hints and modern Python practices
For DevOps Engineers
- Customize DevContainer in
.devcontainer/ - Configure CI/CD pipelines using the Makefile targets
- Deploy documentation using MkDocs build outputs
- Monitor MCP server performance and usage
🔒 Security & Best Practices
- Rootless containers for enhanced security
- No secrets in code - use environment variables
- Input validation in MCP server endpoints
- Type safety with comprehensive type hints
- Dependency scanning with automated security checks
📖 Documentation
Complete documentation is available at:
🤝 Contributing
We welcome contributions! Please see our Contributing Guide for details.
Development Setup
- Fork the repository
- Clone your fork
- Open in DevContainer
- Run
make setup - Make your changes
- Run
make ci-check - Submit a pull request
📋 Requirements
System Requirements
- OS: Linux, macOS, or Windows with WSL2
- RAM: 4GB minimum, 8GB recommended
- Storage: 2GB free space
Software Requirements
- Python: 3.11 or higher
- Podman: Latest stable version
- VSCode: With Dev Containers extension
- Git: For version control
🆘 Troubleshooting
Common Issues
Container build fails
# Clean Podman cache and try again
podman system prune -a
# Or rebuild DevContainer from VSCode
# Ctrl+Shift+P -> "Dev Containers: Rebuild Container"
Port conflicts
make docs-serve MKDOCS_PORT=8002
Dependency issues
make clean
make dev-install
See the Troubleshooting Guide for more solutions.
📄 License
This project is licensed under the MIT License - see the LICENSE file for details.
🙏 Acknowledgments
- MkDocs - Static site generator
- Material for MkDocs - Beautiful theme
- Model Context Protocol - AI integration standard
- uv - Fast Python package manager
- Ruff - Python linting and formatting
- Podman - Container runtime
📞 Support
- Documentation: Project Documentation
- Issues: GitHub Issues
- Discussions: GitHub Discussions
⭐ Star this repo if you find it helpful!
Built with ❤️ using modern Python development practices.
推荐服务器
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。
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。
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。
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。
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。
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。
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。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
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。