RAG + MCP Demo Project

A demonstration project combining RAG (Retrieval-Augmented Generation) with MCP (Model Context Protocol) for building AI agents with local synthetic data.

Project Structure

Project_Rag_Mcp/
├── venv/                  # Virtual environment (created below)
├── data_generator.py      # Synthetic data generation script
├── requirements.txt       # Python dependencies
└── synthetic_data.json    # Generated data (created after running generator)

Setup Instructions

1. Create and Activate Virtual Environment

PowerShell:

# Create virtual environment
python -m venv venv

# Activate virtual environment
.\.venv\Scripts\Activate.ps1

If you get an execution policy error, run:

Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope Process
.\.venv\Scripts\Activate.ps1

2. Install Dependencies

After activating the venv:

pip install --upgrade pip
pip install faker mcp

Or install from requirements file:

pip install -r requirements.txt

3. Generate Synthetic Data

Run the data generator:

python data_generator.py

Output:

• Creates synthetic_data.json with:
  • 20 documents (for RAG retrieval)
  • 10 user profiles
  • 10 search queries

Quick Command Reference

Task	Command
Create venv	python -m venv venv
Activate venv (PS)	.\.venv\Scripts\Activate.ps1
Activate venv (CMD)	venv\Scripts\activate.bat
Install packages	pip install faker mcp
Generate data	python data_generator.py
Deactivate venv	deactivate

Next Steps

• Expand data_generator.py to include embeddings
• Create MCP server for data access
• Build RAG retrieval logic
• Integrate with LLM for generation

Notes

• The faker library generates realistic but random data
• All generated data is local and can be regenerated anytime
• mcp library provides the Model Context Protocol server framework