🤖 Complete Model Context Protocol (MCP) Bootcamp
From Zero to Production — Build Real-World AI Agents with MCP
A hands-on, project-based bootcamp covering everything you need to master the Model Context Protocol — from fundamentals to full-stack production-ready AI agents.
🚀 Get Started • 📚 Modules • 🛠️ Tech Stack • 🤝 Contributing
🌟 What is Model Context Protocol (MCP)?
Model Context Protocol (MCP) is an open standard introduced by Anthropic that defines how AI models communicate with external tools, data sources, and services. Think of it as the "USB standard for AI" — it provides a universal plug-and-play interface that lets any LLM connect to any tool or data source without custom integrations.
Why MCP Matters
📚 Course Modules
This bootcamp is structured as 16 progressive modules, each building on the last — from core concepts to production-ready, multi-agent systems.
📖 01 — Introduction to MCP
Foundation | Conceptual Overview
Understand the architecture of MCP from the ground up. This module covers:
- What MCP is and the problem it solves
- The MCP Client ↔ Server communication model
- Transport layers: stdio vs SSE (Server-Sent Events)
- MCP primitives: Tools, Resources, and Prompts
- How MCP differs from traditional REST API integrations
📄 Includes a comprehensive PDF reference guide.
🖥️ 02 — Build MCP Server with Claude Desktop
Beginner | Python · FastMCP · Claude Desktop
Build your first real MCP server and connect it directly to Claude Desktop. This module walks through:
- Setting up a Python MCP server using
FastMCP - Building a Terminal Server — let Claude execute shell commands on your machine
- Configuring
claude_desktop_config.jsonfor local MCP server registration - Understanding the stdio transport layer in practice
# Your first MCP tool
@mcp.tool()
async def run_command(command: str) -> str:
"""Run a terminal command inside the workspace directory."""
...
📁 terminal_server/ · claude_desktop_config.json
🖱️ 03 — Cursor IDE MCP Server Setup
Beginner | Python · FastMCP · Cursor IDE
Connect your MCP server to Cursor IDE and supercharge your coding workflow:
- Registering MCP servers inside Cursor's settings
- Reusing the terminal server pattern with Cursor's AI assistant
- Understanding how IDEs consume MCP tools natively
- Best practices for scoping tools to development workflows
📁 terminal_server/src/
💎 04 — Build Your Own MCP Client using Google Gemini API
Intermediate | Python · Google Gemini API · Custom MCP Client
Move beyond Claude Desktop — build a fully custom MCP client powered by the Google Gemini API:
- Implementing the MCP client protocol from scratch
- Connecting to any MCP server programmatically
- Integrating Gemini's function calling with MCP tool schemas
- Building a conversational loop that routes tool calls through MCP
- Managing session lifecycle and tool result passing
📁 mcp-client/client.py
🐳 05 — How to Build a Docker MCP Server
Intermediate | Docker · Python · Containerization
Package and deploy your MCP server as a Docker container for portability and reproducibility:
- Writing a production-ready
Dockerfilefor an MCP server - Configuring Claude Desktop to connect to a containerized MCP server
- Managing environment variables and volumes in Docker
- WSL2 setup for Windows users running Docker-based MCP servers
# Containerized MCP server
FROM python:3.11-slim
COPY . /app
RUN pip install -r requirements.txt
CMD ["python", "terminal_server.py"]
📁 terminal_server/ · Dockerfile · claude_desktop_config.json
🦜 06 — LangChain MCP Client using LangChain MCP Adapters
Intermediate | LangChain · Python · MCP Adapters
Integrate MCP tools into the LangChain ecosystem using official MCP adapters:
- Using
langchain-mcp-adaptersto wrap MCP tools as LangChain tools - Building a ReAct agent that reasons over MCP-powered capabilities
- Connecting to multiple MCP servers simultaneously from a LangChain chain
- Streaming agent responses with tool invocations
from langchain_mcp_adapters.tools import load_mcp_tools
tools = await load_mcp_tools(session)
agent = create_react_agent(model, tools)
📁 mcp-clients/lanchain_mcp_client.py
🔗 07 — MCP Client with Multiple Server Support
Intermediate | Python · LangChain · Multi-Server Architecture
Scale your MCP setup by connecting a single client to multiple servers:
- Loading server configurations from a
config.jsonfile - Managing concurrent connections to multiple MCP servers
- Tool namespace management to avoid naming collisions
- Building robust error handling for multi-server environments
- Dynamic server discovery and tool aggregation
{
"servers": {
"terminal": { "command": "python", "args": ["server.py"] },
"weather": { "command": "python", "args": ["weather_server.py"] }
}
}
📁 mcp-client/langchain_mcp_client_wconfig.py · config.json
📡 08 — MCP Server and Client using SSE
Intermediate | Python · SSE Transport · HTTP
Switch from stdio to Server-Sent Events (SSE) transport for network-based MCP communication:
- Building an HTTP-based MCP server with SSE streaming
- Writing an MCP client that connects over HTTP instead of stdio
- When to use SSE vs stdio transport
- Deploying SSE-based MCP servers behind a reverse proxy
- Docker deployment of SSE-based servers
📁 servers/terminal_server/sse_server/ · clients/mcp_client/
🌦️ 09 — Real-Time Weather Agent using MCP and MCP Inspector
Intermediate | Python · FastMCP · External APIs · MCP Inspector
Build a production-style real-time weather agent and debug it with the official MCP Inspector tool:
- Integrating a live weather API as an MCP tool
- Structuring tools in a modular
tools/package - Using the MCP Inspector to visually test and debug your server
- Best practices for API key management in MCP tools
@mcp.tool()
async def get_weather(city: str) -> dict:
"""Fetch real-time weather data for a given city."""
...
📁 tools/weather.py · main.py
💼 10 — Real-Time Job Recommendation System
Advanced | Python · FastMCP · LangChain · Streamlit · Job APIs
A full-stack AI-powered job recommendation agent built on MCP:
- MCP server exposing live job listing APIs as tools
- LangChain agent that understands user profiles and recommends relevant jobs
- Streamlit frontend for an interactive user experience
- Structured output parsing and ranking logic
- Clean separation between MCP server (
mcp_server.py) and agent logic (app.py)
📁 app.py · mcp_server.py · src/job_api.py · src/helper.py
📖 11 — StoryForge Agent
Advanced | Python · FastMCP · LangChain · Streamlit · Generative AI
An end-to-end AI creative writing agent that generates rich, structured stories:
- MCP server with story-generation tools (plot, characters, dialogue)
- Multi-step agent pipeline: outline → draft → refine
- Streamlit web interface for interactive story creation
- Demonstrates long-form content generation with MCP tools
- Extensible tool design for adding new creative capabilities
📁 app.py · mcp_server.py · main.py
🏥 12 — Clinisight — Clinical AI Assistant
Advanced | Python · FastMCP · PubMed API · LangChain · Medical NLP
A specialized clinical intelligence agent for medical research and symptom analysis:
- MCP tools for PubMed article retrieval and summarization
- Symptom extraction from natural language clinical notes
- AI-powered differential diagnosis suggestions
- Integration with biomedical literature APIs
- Jupyter notebook for exploratory clinical data analysis
Tools: symptom_extractor · pubmed_search · summarize_article · differential_diagnosis
📁 functions/ · mcp_tool.py · app.py · demo.ipynb
🤖 13 — ADK Multi-Agent Demo
Advanced | Google ADK · Multi-Agent Systems
Explore Google's Agent Development Kit (ADK) for orchestrating multiple specialized agents:
- Building a multi-agent system where agents collaborate on tasks
- Agent-to-agent communication patterns
- Orchestration strategies: sequential, parallel, and hierarchical
- Integrating ADK agents with external MCP tools
📁 muli_agent/agent.py
✈️ 14 — TripMate-AI
Advanced | Python · Flask · LangChain · Flight APIs · Tavily Search
A full-stack AI travel planning assistant with a beautiful web interface:
- Flask web application with custom HTML/CSS/JS frontend
- Real-time flight search integration
- Tavily search tools for destination research
- Multi-turn conversational travel planning
- Dockerized for easy deployment
📁 app.py · backend.py · tools/ · templates/ · static/ · Dockerfile
✈️🔌 15 — TripMate-AI Using MCP
Advanced | Python · MCP · Flask · Custom Weather MCP Server
The MCP-powered evolution of TripMate-AI — refactored to use MCP as the tool integration layer:
- Custom Weather MCP Server (
custom_weather_mcp_server.py) - MCP Client managing server connections from the backend
- Decoupled architecture: frontend → Flask → MCP Client → MCP Servers
- Demonstrates real-world migration from direct API calls to MCP
- Full Docker support for production deployment
Architecture: Browser → Flask App → MCP Client → [Weather MCP Server, Flight Tools, Search Tools]
📁 mcp_client.py · custom_weather_mcp_server.py · backend.py · Dockerfile
✍️ 16 — AgentWriter-AI
Advanced | Python · LangChain · Google Gemini · DALL-E · Flask
A sophisticated AI blog writing agent with image generation capabilities:
- Multi-stage writing pipeline: research → outline → draft → polish
- DALL-E image generation integrated into blog posts
- Web scraping and Tavily research for factual grounding
- Beautiful Flask web frontend for article creation
- Jupyter notebooks documenting each stage of agent evolution
Pipeline: Topic → Research → Outline → Draft → Image Generation → Final Blog Post
📁 app.py · backend.py · notebooks/ · static/ · templates/
🛠️ Tech Stack
🚀 Getting Started
Prerequisites
Make sure you have the following installed:
- Python 3.11+
- Git
- Docker (for containerized projects)
uv(recommended) orpip
Clone the Repository
git clone https://github.com/Suraj-G-Rao/Complete-Model-Context-Protocol-MCP-.git
cd Complete-Model-Context-Protocol-MCP-
Setting Up a Project
Each module is self-contained. Navigate to any folder and follow these steps:
Using uv (recommended):
cd "04-Build Your Own MCP Client using Google Gemini API/mcp-client"
uv sync
cp .env.example .env # Add your API keys
uv run python client.py
Using pip:
cd "04-Build Your Own MCP Client using Google Gemini API/mcp-client"
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env # Add your API keys
python client.py
Environment Variables
Each project that requires API keys includes a .env.example file. Common variables:
GOOGLE_API_KEY=your_google_gemini_api_key
ANTHROPIC_API_KEY=your_anthropic_api_key
OPENAI_API_KEY=your_openai_api_key
TAVILY_API_KEY=your_tavily_api_key
OPENWEATHER_API_KEY=your_openweather_api_key
📂 Repository Structure
Complete-Model-Context-Protocol-MCP-/
│
├── 01-Introduction to MCP/ # 📖 MCP concepts & PDF guide
├── 02-Build MCP Server with Claude Desktop/ # 🖥️ First MCP server + Claude Desktop
├── 03-Cursor IDE MCP Server Setup/ # 🖱️ MCP in Cursor IDE
├── 04-Build Your Own MCP Client.../ # 💎 Custom Gemini-powered MCP client
├── 05-How to build Docker MCP Server/ # 🐳 Dockerized MCP server
├── 06-Langchain MCP Client.../ # 🦜 LangChain + MCP adapters
├── 07-MCP Client with Multiple Server.../# 🔗 Multi-server MCP client
├── 08-MCP Server and Client using SSE/ # 📡 SSE transport layer
├── 09-Real Time Weather Agent.../ # 🌦️ Weather agent + MCP Inspector
├── 10-Real Time Job Recommendation.../ # 💼 Full-stack job recommender
├── 11-StoryForge Agent/ # 📖 AI creative writing agent
├── 12-Clinisight/ # 🏥 Clinical AI assistant
├── 13-ADK demo/ # 🤖 Google ADK multi-agent
├── 14-TripMate-AI/ # ✈️ Travel planner web app
├── 15-TripMate-AI Using MCP/ # ✈️🔌 MCP-powered travel planner
├── 16-AgentWriter-AI/ # ✍️ AI blog writing agent
│
├── .gitignore
└── README.md
🗺️ Learning Path
Beginner ─────────────────────────────────────────────────────► Advanced
│ │
[01] [02] [03] [04] [05] [06] [07] [08] │
Concepts First Server Custom Client LangChain & Docker │
│ │ │ │
[09] [10] [11] [12] [13] [14][15][16]
Weather Agent Job Rec. Full-Stack Agents Production Apps
🤝 Contributing
Contributions are welcome! If you find a bug, want to improve documentation, or add a new module:
- Fork the repository
- Create a feature branch:
git checkout -b feature/your-feature - Make your changes and commit:
git commit -m "Add your feature" - Push to your branch:
git push origin feature/your-feature - Open a Pull Request
📜 License
This project is licensed under the MIT License — see the LICENSE file for details.
⭐ If this bootcamp helped you, give it a star!
Built with ❤️ to help developers master the future of AI tool integration
