📄 SmartDoc — AI Document Q&A

SmartDoc is an AI-powered document assistant that lets you upload PDF, DOCX, or TXT files and ask questions in natural language. Built with FastAPI, LangChain, OpenAI, and React.

🚀 Live Demo

• Frontend: smartdoc.imranansari.in
• API Docs: api.smartdoc.imranansari.in/docs

---

✨ Features

• 📁 Upload PDF, DOCX, and TXT files (max 2MB)
• 🤖 Ask questions in natural language
• 🧠 Smart context detection — uses RAG for specific questions, full document for broad questions (summarize, review, describe)
• 💬 Multi-turn chat history per session
• 🔒 IP-based rate limiting — 3 uploads and 12 questions per day
• 🐳 Dockerized for easy deployment
• ⚡ Redis-backed rate limiting with auto daily reset

---

🛠️ Tech Stack

Backend

Technology	Purpose
FastAPI	REST API framework
LangChain	RAG pipeline and LLM orchestration
OpenAI GPT-4o-mini	Language model for Q&A
OpenAI text-embedding-3-small	Document embeddings
FAISS	Vector store for semantic search
Redis	Rate limiting storage
Docker	Containerization

Frontend

Technology	Purpose
React + Vite	UI framework
Axios	HTTP client
React Markdown	Markdown rendering

---

🏗️ Architecture

User uploads file
      ↓
Text extracted (PDF / DOCX / TXT)
      ↓
Text chunked → Embeddings generated → Stored in FAISS
      ↓
User asks question
      ↓
Smart routing:
  Specific question → RAG (top 4 relevant chunks)
  Broad question    → Full document context
      ↓
GPT-4o-mini generates answer
      ↓
Response shown in chat UI

---

📁 Project Structure

smartdoc/
├── backend/
│   ├── main.py                 # FastAPI app entry point
│   ├── limiter.py              # Redis rate limiting
│   ├── requirements.txt
│   ├── Dockerfile
│   ├── .env.example
│   ├── routers/
│   │   ├── upload.py           # File upload endpoint
│   │   └── chat.py             # Q&A endpoint
│   └── services/
│       ├── extractor.py        # PDF / DOCX / TXT parser
│       ├── embedder.py         # Chunking + FAISS vector store
│       └── rag.py              # LangChain RAG chain
├── frontend/
│   ├── src/
│   │   ├── App.jsx
│   │   ├── App.css
│   │   └── components/
│   │       ├── Uploader.jsx    # File upload UI
│   │       └── ChatBox.jsx     # Chat interface
│   └── .env.example
|── docker-compose.yml
|
|__ makefile

---

⚙️ Local Setup

Prerequisites

• Python 3.11+
• Node.js 18+
• Redis
• OpenAI API key

Backend

cd backend

# Create virtual environment
python3.11 -m venv .venv
source .venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Configure environment
cp .env.example .env
# Add your OPENAI_API_KEY in .env

# Run server
uvicorn main:app --reload

Backend runs at http://127.0.0.1:8000 API docs at http://127.0.0.1:8000/docs

Frontend

cd frontend

# Install dependencies
npm install

# Configure environment
cp .env.example .env
# Set VITE_API_URL=http://127.0.0.1:8000

# Run dev server
npm run dev

Frontend runs at http://localhost:5173

---

🐳 Docker Setup

# Start all services
docker-compose up -d

# View logs
docker logs smartdoc-backend

# Stop all services
docker-compose down

---

🔌 API Endpoints

Method	Endpoint	Description
GET	/	Health check
POST	/api/upload	Upload a document
POST	/api/chat	Ask a question
POST	/api/clear	Clear session history

Upload a file

curl -X POST https://api.smartdoc.imranansari.in/api/upload \
  -F "file=@document.pdf"

Response:

{
  "session_id": "550e8400-e29b-41d4-a716-446655440000",
  "filename": "document.pdf",
  "message": "File uploaded and processed successfully"
}

Ask a question

curl -X POST https://api.smartdoc.imranansari.in/api/chat \
  -H "Content-Type: application/json" \
  -d '{
    "session_id": "550e8400-e29b-41d4-a716-446655440000",
    "question": "What is this document about?"
  }'

Response:

{
  "session_id": "550e8400-e29b-41d4-a716-446655440000",
  "question": "What is this document about?",
  "answer": "..."
}

---

🔒 Rate Limits

Action	Limit
File uploads	3 per day per IP
Questions	12 per day per IP

Limits reset automatically every 24 hours.

---

🌍 Environment Variables

Backend .env

OPENAI_API_KEY=your_openai_api_key_here

Frontend .env

VITE_API_URL=http://127.0.0.1:8000

---

👨‍💻 Author

Imran Ansari

• Portfolio: imranansari.in
• Blog: blog.imranansari.in
• GitHub: github.com/imranansari

---

📄 License

MIT License — feel free to use this project for learning or portfolio purposes.