AI导哈

MCP学习资源库

一个包含MCP协议实现、示例与学习资源的开源仓库，帮助开发者系统学习并构建上下文感知的AI应用。

  • by Mohamed-Noufal
  • Python
  • 本地部署
  • 更新于
在 GitHub 查看 Mohamed-Noufal/MCP

使用说明

MCP (Model Context Protocol) Learning Repository

This repository contains a comprehensive collection of MCP (Model Context Protocol) implementations, examples, and projects from various learning resources, primarily focused on building rich-context AI applications with Anthropic and other AI frameworks.

📚 Repository Structure

Core Learning Materials/

Crash Course

Complete MCP crash course covering fundamentals to advanced topics:

  • Introduction and Context
  • Understanding MCP
  • Simple Server Setup (SSE, STDIO, Streamable HTTP)
  • OpenAI Integration
  • MCP vs Function Calling
  • Docker Deployment
  • Lifecycle Management

Deep Learning Course Lessons

Progressive lessons from the Anthropic MCP course:

  • L3: Chatbot Example
  • L4: Creating an MCP Server
  • L5: Creating an MCP Client
  • L6: Connecting MCP Chatbot to Reference Servers
  • L7: Adding Prompt & Resource Features
  • L8: Creating and Deploying Remote Servers

Project Implementations

  • 📝 Notion-MCP-Agent: Integration between Notion and MCP
  • 📧 mcp-notion-email: Email integration with Notion via MCP
  • 📄 Docs-sheet-mcp: Google Sheets/Docs MCP integration
  • 🌐 Community Servers: Various MCP server implementations
  • 🧪 Acadamic-Research: Research and experimental implementations

Additional Resources

  • Appendx: Supporting materials, images, and documentation
  • 1/: Simple agent implementation examples

🚀 Getting Started

Prerequisites

  • Python 3.13 or higher
  • Node.js (for JavaScript-based servers)
  • Docker (optional, for containerized deployments)
  • Git

Installation

  1. Clone the repository:
git clone https://github.com/Mohamed-Noufal/MCP.git
cd N-MCP
  1. Create a virtual environment:
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
  1. Install dependencies for specific projects:
# For Python projects
pip install -r requirements.txt

# For Node.js projects (e.g., Notion-server-)
cd Notion-server-
npm install

Environment Setup

Create a .env file based on .env.example (found in crash-course) and add your API keys:

OPENAI_API_KEY=your_key_here
ANTHROPIC_API_KEY=your_key_here
# Add other required keys

📖 How to Use This Repository

For Learning

Navigate through the lessons in order (L3 → L8) to understand MCP progressively. Each lesson contains:

  • Jupyter notebooks with explanations
  • Python implementation files
  • README documentation
  • Supporting images and resources

For Reference

Browse the crash-course directory for quick references on specific MCP topics and implementations.

For Development

Use the various server implementations in Community Servers, Notion-server-, and other project directories as templates for your own MCP servers.

🔧 Key Technologies

  • MCP (Model Context Protocol): Core protocol for AI context management
  • Python: Primary implementation language
  • OpenAI API: LLM integration
  • Anthropic Claude: Advanced AI model integration
  • Docker: Containerization and deployment
  • Notion API: Notion workspace integration
  • Google APIs: Sheets/Docs integration

📂 Project Highlights

MCP Chatbot Client

Found in L5-L7, demonstrates how to:

  • Create MCP clients
  • Connect to reference servers
  • Implement prompt and resource features

Research Server

Custom MCP server implementation showcasing:

  • Tool definition and registration
  • Resource management
  • Prompt templates

Notion Integration

Full-featured Notion MCP server with:

  • Email integration
  • Database operations
  • Agent-based interactions

🤝 Contributing

This is a learning repository. Feel free to:

  • Explore the code
  • Try the examples
  • Create your own MCP implementations
  • Share improvements

📝 License

Please refer to individual project directories for specific licensing information.

🔗 Resources

📧 Contact

For questions or collaboration:

Note: This repository is for educational purposes. Ensure you have proper API keys and credentials before running any examples that require external services.

相关 MCP