Model Context Protocol (MCP)

Model Context Protocol (MCP) is an open protocol that standardizes how applications provide context to Large Language Models (LLMs). Think of MCP like a USB-C port for AI applications - providing a standardized way to connect AI models to different data sources and tools.

• Standardized Integration: Connect LLMs to various data sources and tools using a common protocol
• Tool Execution: Enable LLMs to perform actions through your server
• Resources Access: Expose data and content from your servers to LLMs
• Prompts Management: Create reusable prompt templates and workflows
• Sampling: Allow servers to request completions from LLMs
• Secure Communication: Built-in transport mechanisms for reliable data exchange

• MCP Host: Programs like Claude Desktop, IDEs, or AI tools that want to access data through MCP
• MCP Client: Protocol clients that maintain 1:1 connections with servers
• MCP Server: Lightweight programs that expose specific capabilities through the standardized Model Context Protocol

MCP includes two standard transport implementations:

• Enables communication through standard input and output streams
• Useful for local integrations and command-line tools
• Ideal for: building command-line tools, implementing local integrations, simple process communication, working with shell scripts

• Enables server-to-client streaming with HTTP POST requests for client-to-server communication
• Useful when: only server-to-client streaming is needed, working with restricted networks, implementing simple updates
• Security Warning: SSE transports can be vulnerable to DNS rebinding attacks if not properly secured

1. Create project folder:

uv init mcpcrashcourse

2. Create server directory and files:

mkdir server
touch server/weather.py

3. Create virtualenv:

uv venv

4. Install MCP dependencies:

uv add "mcp[cli]"

5. Run your server:

uv run mcp dev server/weather.py

uv run mcp install server/weather.py

Create a client file and integrate using libraries like langchain-groq.

A Dockerfile is provided to containerize your MCP server.

• Official Documentation: https://modelcontextprotocol.io/introduction
• Anthropic Blog: https://www.anthropic.com/news/model-context-protocol
• Official Servers: https://github.com/modelcontextprotocol/servers
• Python SDK: https://github.com/modelcontextprotocol/python-sdk
• VSCode Integration: https://code.visualstudio.com/docs/copilot/chat/mcp-servers

• Smithery (3rd party MCP Servers): https://smithery.ai/
• MCP-Use: https://github.com/mcp-use/mcp-use