🧠 DocMind — RAG-Powered Document Q&A

An end-to-end Retrieval-Augmented Generation (RAG) chatbot that lets you have a conversation with any PDF document — powered by Google Gemini AI and semantic vector search.

---

✨ Features

• 📄 Upload any PDF — research papers, books, reports, manuals
• 🔍 Semantic search using FAISS vector index with cosine similarity
• 🤖 Gemini 1.5 Flash for fast, intelligent, context-aware answers
• 📌 Page-number citations on every response
• 💬 Multi-turn memory — the AI remembers your conversation
• ⚡ Local embeddings with all-MiniLM-L6-v2 (no extra API cost)
• 🎨 Premium dark UI with glassmorphism styling

---

🚀 Quick Start

1. Clone the repo

git clone https://github.com/YOUR_USERNAME/docmind-rag-chatbot.git
cd docmind-rag-chatbot

2. Create a virtual environment (recommended)

python -m venv venv
venv\Scripts\activate      # Windows
# source venv/bin/activate  # macOS/Linux

3. Install dependencies

pip install -r requirements.txt

4. Set up your API key

Option A — UI (easiest): Just paste your key in the sidebar when the app opens.

Option B — .env file (recommended for local use):

cp .env.example .env
# Then open .env and add your key

GOOGLE_API_KEY=AIzaSy...your_key_here

Get a free key at 👉 aistudio.google.com

5. Run the app

streamlit run app.py

---

🏗️ How It Works

PDF Upload
    │
    ▼
PyMuPDF → Extract text (page by page)
    │
    ▼
Chunker → 500-word overlapping chunks
    │
    ▼
SentenceTransformer → Dense vector embeddings
    │
    ▼
FAISS Index → Store & search vectors
    │
    ▼
User Query → Embed → Retrieve top-5 chunks
    │
    ▼
Gemini 1.5 Flash → Generate cited answer

---

🛠️ Tech Stack

Layer	Tool
LLM	Google Gemini 1.5 Flash
Embeddings	sentence-transformers (all-MiniLM-L6-v2)
Vector Store	FAISS (cosine similarity)
PDF Parsing	PyMuPDF (fitz)
UI	Streamlit
Env Management	python-dotenv

---

📁 Project Structure

docmind-rag-chatbot/
├── app.py              # Main Streamlit application
├── requirements.txt    # Python dependencies
├── .env.example        # API key template
├── .gitignore          # Git ignore rules
└── README.md           # This file

---

⚠️ Security Note

• Never commit your .env file — it's already in .gitignore
• The sidebar API key input is designed for sharing the app; for personal use, prefer the .env approach

---

📄 License

MIT License — feel free to use, modify, and share.