DocFlow - AI Document Assistant (RAG System)

DocFlow is a full-stack, AI-powered document assistant that lets users upload PDFs and chat with their content using RAG (Retrieval-Augmented Generation).
It uses Hugging Face Inference API, FAISS vector search, FastAPI, and Next.js to deliver powerful document understanding and question answering.

---

Features

• Upload any PDF document
• Extract & chunk text automatically
• Semantic search using FAISS
• LLM-powered answers using HuggingFace models
• RAG pipeline: retrieve relevant chunks & generate contextual answers
• Chat interface built with Next.js
• Fast, modern UI using Tailwind CSS
• Local processing + secure API communication

---

Tech Stack

Backend

• Python 3.10+
• FastAPI
• FAISS (Vector Database)
• HuggingFace Inference API
• PyPDF2
• Uvicorn

Frontend

• Next.js 16
• React 19
• Tailwind CSS 4
• TypeScript

---

Project Structure

DocFlow-AI/
├── backend/
│   ├── rag_app.py          # Main FastAPI application
│   ├── .env               # Environment variables (HF_TOKEN)
│   ├── venv/              # Python virtual environment
│   └── __pycache__/       # Python cache files
├── frontend/
│   ├── app/
│   │   ├── globals.css    # Global styles with Tailwind
│   │   ├── layout.tsx     # Root layout component
│   │   ├── page.tsx       # Main page component
│   │   └── favicon.ico    # App favicon
│   ├── public/            # Static assets
│   ├── node_modules/      # Node.js dependencies
│   ├── package.json       # Node.js dependencies and scripts
│   ├── tailwind.config.js # Tailwind CSS configuration
│   ├── postcss.config.mjs # PostCSS configuration
│   ├── next.config.ts     # Next.js configuration
│   ├── tsconfig.json      # TypeScript configuration
│   ├── eslint.config.mjs  # ESLint configuration
│   └── .gitignore         # Git ignore rules
├── .git/                  # Git repository
└── README.md             # Project documentation

---

Environment Variables

Create a .env file inside the backend folder:

HF_TOKEN=your_huggingface_api_key_here

Installation & Setup

Clone the repository

git clone https://github.com/Malindup2/DocFlow-AI.git
cd DocFlow-AI

Backend Setup (FastAPI + FAISS)

cd backend
python -m venv venv
venv\Scripts\activate   # Windows
# OR
source venv/bin/activate  # Mac/Linux
pip install fastapi uvicorn[standard] faiss-cpu huggingface_hub pypdf2 python-multipart python-dotenv
uvicorn rag_app:app --reload

Backend will run at: http://127.0.0.1:8000

Test using Swagger Docs: http://127.0.0.1:8000/docs

Frontend Setup (Next.js + Tailwind)

cd ../frontend
npm install
npm run dev

Frontend runs at: http://localhost:3000