DocuGen

DocuGen is a MERN-stack document and PDF utility suite. It lets users generate invoices, convert between images and PDFs, extract text via OCR, compress files, build reusable document templates, manage Google Drive files, and chat with uploaded documents through a Gemini-powered RAG assistant.

Features

• JWT-based signup / login
• Smart invoice generator with GST calculation, PDF preview, and download
• Image → PDF conversion (batch)
• PDF → image (JPEG) conversion
• OCR text extraction (Tesseract.js)
• Image and PDF compression
• Custom document template builder (fields, layout, styling) with a template library and sharing
• Google Drive integration — list, upload, rename, delete files
• AI document assistant — upload a PDF/DOCX/TXT, it's chunked and embedded, then answered against via Gemini (retrieval-augmented generation)

Tech Stack

Backend

• Node.js, Express 4
• MongoDB, Mongoose
• JWT (jsonwebtoken) + bcryptjs for auth
• Multer for uploads
• PDFKit, pdf-lib, Puppeteer, pdf-poppler, pdf2pic for PDF generation/conversion
• Tesseract.js for OCR
• Sharp for image processing
• googleapis for Google OAuth2 + Drive
• Cloudinary (media storage)
• vectra for the local vector store (RAG embeddings)
• Helmet, express-rate-limit, express-validator (installed; not yet wired into any routes — see Known Issues)

Frontend

• React 19, Vite 7
• React Router v7
• Axios
• lucide-react (icons)
• Tailwind CSS (installed; most of the UI currently uses inline styles rather than Tailwind classes)

Project Structure

docugen/
├── backend/
│   ├── controller/
│   │   └── ocrController.js
│   ├── middleware/
│   │   └── authMiddleware.js        # JWT verification
│   ├── models/
│   │   ├── user.js
│   │   ├── Template.js
│   │   ├── GeneratedDocument.js
│   │   ├── InvoiceSchema.js
│   │   └── Document.js
│   ├── routes/
│   │   ├── auth.js                  # signup / login / me
│   │   ├── googleAuth.js            # Google OAuth connect + callback
│   │   ├── invoiceRoute.js
│   │   ├── pdfRoutes.js             # images -> PDF
│   │   ├── pdfToImageRouteMain.js   # PDF -> images
│   │   ├── ocrRoute.js
│   │   ├── compressRoutes.js
│   │   ├── driveRoutes.js
│   │   ├── templateRoutes.js
│   │   └── aiAssistantRoute.js      # RAG assistant (Gemini)
│   ├── templates/                   # server-side HTML/PDF templates
│   ├── utils/                       # cloudinary, PDF gen, Google Drive client, multer, route loader
│   ├── vector_store/                # local embeddings cache (gitignored)
│   └── server.js
│
└── frontend/
    └── src/
        ├── Context/ToastContext.jsx
        ├── pages/                   # one file per route/feature
        ├── services/api.js          # axios instance, injects JWT from localStorage
        ├── templates/                # client-side invoice template
        └── App.jsx                  # top-level router

Getting Started

Clone the repository:

git clone https://github.com/your-username/docugen.git
cd docugen

Install dependencies:

cd backend && npm install
cd ../frontend && npm install

Set up backend environment variables:

cd backend
cp .env.example .env

Fill in .env:

PORT=5000
NODE_ENV=development
MONGO_URI=mongodb+srv://<user>:<password>@cluster0.xxxxx.mongodb.net/DocuGen?retryWrites=true&w=majority
JWT_SECRET=<generate with: openssl rand -hex 32>
JWT_EXPIRE=1d
GOOGLE_CLIENT_ID=your_google_client_id
GOOGLE_CLIENT_SECRET=your_google_client_secret
GOOGLE_REDIRECT_URI=http://localhost:5000/api/auth/google/callback
FRONTEND_URL=http://localhost:5173
GEMINI_API_KEY=your_gemini_api_key
MAX_FILE_SIZE=10485760
UPLOAD_DIR=uploads/

Set up frontend environment variables (frontend/.env):

VITE_API_URL=http://localhost:5000/api

Run both apps in development:

# backend
cd backend && npm run dev

# frontend (separate terminal)
cd frontend && npm run dev

Backend runs on http://localhost:5000, frontend on http://localhost:5173.

API Overview

Prefix	Routes file	Notes
/api/auth	auth.js, googleAuth.js	signup, login, me, Google OAuth
/api/invoice	invoiceRoute.js	generate, preview, download, history
/api/pdf	pdfRoutes.js	images → PDF
/api/pdf-to-image	pdfToImageRouteMain.js	PDF → images
/api/ocr	ocrRoute.js	text extraction
/api/compress	compressRoutes.js	image/PDF compression
/api/drive	driveRoutes.js	Drive file management
/api	templateRoutes.js	template CRUD, sharing, generation
/api/ai	aiAssistantRoute.js	RAG upload/ask/status/clear
/health	server.js	health check

All routes except /api/auth/signup, /api/auth/login, and /health require a Bearer <token> header.

Deployment

• Frontend → Vercel (static Vite build)
• Backend → Railway / Render (Puppeteer and local file writes mean it cannot run on Vercel's serverless functions)

Set VITE_API_URL on the frontend host and FRONTEND_URL + the Google OAuth redirect URI on the backend host to match your deployed domains.