---
slug: "ragflow"
source_type: "readme"
source_url: "https://cdn.jsdelivr.net/gh/hulk-pham/MCP-RAG@main/README.md"
repo: "https://github.com/hulk-pham/MCP-RAG"
source_file: "README.md"
branch: "main"
---
# RAG Application
A demo of Retrieval-Augmented Generation (RAG) application with MCP server integration.

![Screenshot](https://github.com/hulk-pham/MCP-RAG/raw/HEAD/github/screenshot.png)

## Features
- MCP server integration
- Document retrieval using vector search with ChromaDB
- Context-aware prompt generation
- Integration with LLM APIs

## Installation
```
pip install -r requirements.txt
```

## Usage
Connect to the MCP server with Claude Desktop, Cursor, or your preferred IDE.

Use the `process_query` tool to ask questions about the company.

## Configuration
Set up your environment variables in .env:
```
OPENAI_API_KEY=your_api_key
```

## Project Structure
app/retrieval.py: Document retrieval functionality
app/context.py: Context management
app/llm_client.py: LLM API integration
app/prompt_builder.py: Prompt construction

## License
MIT