LLM_MCP_service_scratch

🚀 MCP Zero to One: Mastering Model Context Protocol Development    Welcome to the **"Mastering MCP from Scratch"** course project! This repository contains comprehensive teaching materials, hands-on exercises, and example code to help developers quickly master Model Context Protocol (MCP) development techniques. ## 🎯 Course Objectives - **Understand Core Concepts**: Gain in-depth understanding of LLM limitations and how MCP solves context integration issues through standardized protocols. - **Master Implementation Skills**: Learn to develop Tools, Resources, and Prompts using Python SDK. - **Ecosystem Integration**: Integrate your MCP Server with Claude Desktop or custom Agent applications. - **Advanced Applications**: Explore Docker deployment, Remote Server, and security best practices. ## 📂 Chapter Navigation | Chapter | Title | Content Introduction | Key Resources | | :--- | :--- | :--- | :--- | | **Ch01** | [Why MCP](./01_why_mcp/) | LLM Limitations and Agent Architecture | [Slides](./01_why_mcp/slides.md), [Agent Loop](./01_why_mcp/diagrams/agent_loop.mermaid) | | **Ch02** | [What is MCP](./02_what_is_mcp/) | Protocol Architecture and Before/After Comparison | [Slides](./02_what_is_mcp/slides.md), [Comparison](./02_what_is_mcp/before_after_comparison.md) | | **Ch03** | [Core Concepts](./03_core_concepts/) | Tools, Resources, Prompts Explained | [Slides](./03_core_concepts/slides.md), [Schema](./03_core_concepts/schema_samples/tool_schema.json) | | **Ch04** | [Protocol Basics](./04_protocol_basics/) | JSON-RPC, stdio, SSE Principles | [Slides](./04_protocol_basics/slides.md), [Sequence](./04_protocol_basics/sequence_charts/tool_execution.mermaid) | | **Ch05** | [Implementing Your First Server](./05_first_mcp_server/) | **🔥 Core Implementation**: Hello World and Calculator | [Code: Basic](./05_first_mcp_server/labs/basic_tool.py), [Guide](./05_first_mcp_server/guides/inspector_guide.md) | | **Ch06** | [Client App Integration](./06_mcp_clients/) | Claude Desktop Setup and Python Client | [Config](./06_mcp_clients/configs/claude_desktop_config.json), [Code: Client](./06_mcp_clients/labs/client.py) | | **Ch07** | [Remote Advanced Applications](./07_remote_advanced/) | Docker Containerized Deployment (SSE) | [Code: Remote](./07_remote_advanced/labs/remote_server.py), [Docker Guide](./07_remote_advanced/guides/docker_deploy.md) | | **Ch08** | [Security and Future](./08_security_future/) | Security Checklist and Ecosystem Outlook | [Checklist](./08_security_future/security_checklist.md), [Slides](./08_security_future/slides.md) | | **Ch09** | [Final Project](./09_final_project/) | Building Your Personal AI Assistant Toolbox | [Project Brief](./09_final_project/project_brief.md) | ## 🛠️ Quick Start ### 1. Environment Preparation This project uses `uv` for package management (10-100 times faster than pip). ```bash # Install uv (if not installed) curl -LsSf https://astral.sh/uv/install.sh | sh # Clone the project git clone https://github.com/Zenobia000/LLM_MCP_service_scratch.git cd mcp-zero-to-one-course # Initialize environment uv sync ``` ### 2. Launch Your First Server (Inspector Mode) ```bash uv run mcp dev 05_first_mcp_server/labs/basic_tool.py ``` ### 3. Deploy Remote Server with Docker ```bash cd 07_remote_advanced/labs docker build -t mcp-remote-server . docker run -d -p 8090:8000 --name my-mcp-server mcp-remote-server ``` ## 📚 Reference Resources - [Model Context Protocol Official Documentation](https://modelcontextprotocol.io/) - [MCP Servers Gallery](https://github.com/modelcontextprotocol/servers) --- *Maintained by MCP Course Team. Last Updated: 2026-01-15