🤖 RAGDoc Assistant

An AI-powered Retrieval-Augmented Generation (RAG) Document Assistant that allows users to upload PDF, DOCX, and TXT files and ask questions in natural language. The application retrieves relevant document content using semantic search and generates accurate answers with Google Gemini AI while displaying reference sources.

---

🌐 Live Application: https://ragdoc-assistant.onrender.com/

---

📸 Screenshots

🏠 Landing Page

---

💬 Main Chat Interface

---

✨ Features

• 📄 Upload PDF, DOCX and TXT documents
• 🤖 AI-powered question answering using Google Gemini
• 🔍 Semantic search with vector embeddings
• 🧠 Retrieval-Augmented Generation (RAG)
• ☁️ Supabase Vector Database integration
• 📚 Automatic document chunking
• 📌 Source references with page numbers
• 🗑️ Clear indexed documents
• 🎨 Modern and responsive user interface

---

🛠️ Tech Stack

Frontend

• HTML5
• CSS3
• JavaScript

Backend

• Python
• Flask

AI & Machine Learning

• Google Gemini 2.5 Flash
• Gemini Embedding Model
• LangChain

Database

• Supabase
• pgvector

Document Processing

• PyPDF
• python-docx

---

📂 Project Structure

RAGDoc_Assistant/
│
├── app.py
├── requirements.txt
├── README.md
├── .gitignore
│
├── data/
├── static/
├── templates/
│
├── landing-page.png
├── main-ui.png
│
└── .env 

---

⚙️ Installation

Clone Repository

git clone https://github.com/yourusername/RAGDoc-Assistant.git

cd RAGDoc-Assistant

Create Virtual Environment

python -m venv .venv

Activate

Windows

.venv\Scripts\activate

Install Dependencies

pip install -r requirements.txt

---

🔑 Environment Variables

Create a .env file in the project root.

GEMINI_API_KEY=YOUR_GEMINI_API_KEY

SUPABASE_URL=YOUR_SUPABASE_URL

SUPABASE_KEY=YOUR_SUPABASE_KEY

LLM_PROVIDER=gemini

GEMINI_QA_MODEL=models/gemini-2.5-flash

GEMINI_EMBEDDING_MODEL=models/gemini-embedding-2

FLASK_SECRET_KEY=your_secret_key

---

▶️ Run the Project

python app.py

Open your browser and visit:

http://127.0.0.1:5000

---

🔄 RAG Workflow

Upload Document
        │
        ▼
Extract Text
        │
        ▼
Text Chunking
        │
        ▼
Gemini Embeddings
        │
        ▼
Supabase Vector Database
        │
        ▼
User Question
        │
        ▼
Similarity Search
        │
        ▼
Relevant Context
        │
        ▼
Gemini LLM
        │
        ▼
Final Answer with Source References

---

📦 Dependencies

• Flask
• LangChain
• Google Generative AI
• Supabase
• PyPDF
• python-docx

---

🔒 Security

• Environment variables are stored in a .env file.
• API keys are never committed to GitHub.
• .env is ignored using .gitignore.

---

🚀 Future Enhancements

• Multiple document collections
• OCR support for scanned PDFs
• User authentication
• Chat history
• Dark mode
• Cloud deployment improvements
• Mobile responsive enhancements

---

👨‍💻 Author : Yatendra Kumar Gupta

---

⭐ Support

If you found this project helpful, please consider giving it a ⭐ Star on GitHub!# RAGDoc_Assistant