PDF Seal Application

An AI-powered web application for applying company seals to PDF documents with intelligent placement suggestions.

Features

• 🤖 AI-Powered Placement: Uses Azure OpenAI GPT-4.1 vision to analyze documents and suggest optimal seal positions
• 📄 Universal Document Support: Works with any document type (contracts, invoices, letters, certificates, forms)
• ✨ Live Preview: Real-time PDF preview with seal overlay
• 🎯 Manual Controls: Fine-tune seal position, size, and page placement
• 🔒 Secure Access: Password-protected with JWT authentication
• 📥 Easy Download: Apply seal and download modified PDF instantly

Getting Started

Prerequisites

• Node.js 20+
• npm or yarn
• Azure OpenAI account with GPT-4.1 deployment

Installation

1. Clone the repository:

git clone <repository-url>
cd pdf-seal-app

2. Install dependencies:

npm install

3. Create .env.local file with required variables:

AZURE_OPENAI_ENDPOINT=https://your-resource.openai.azure.com
AZURE_OPENAI_API_KEY=your-api-key
AZURE_OPENAI_DEPLOYMENT=gpt-4.1
MASTER_PASSWORD=your-password
JWT_SECRET=your-jwt-secret

4. Run the development server:

npm run dev

5. Open http://localhost:3000 in your browser

Build for Production

npm run build
npm start

How It Works

1. Upload PDF: Drag and drop or select a PDF document
2. AI Analysis: Click "AI Suggest Position" to get intelligent placement recommendations
3. Manual Adjustment: Fine-tune position, size, and page using controls
4. Preview: See real-time preview of seal placement
5. Apply & Download: Generate and download the sealed PDF

Technology Stack

• Framework: Next.js 15 with Turbopack
• UI: React 19, Tailwind CSS 4, Radix UI
• PDF Processing: pdf-lib, Puppeteer, react-pdf
• AI: Azure OpenAI GPT-4.1 Vision
• Authentication: JWT with jose library
• Image Processing: Sharp, Canvas

Architecture Highlights

Vision-Based AI Analysis

• Converts PDF pages to high-resolution images using Puppeteer (headless Chrome)
• Sends images to GPT-4.1 vision API for intelligent analysis
• Generalized prompts work with any document type without hardcoded assumptions
• Returns coordinates, reasoning, and confidence score

Pure npm Solution

• No system-level dependencies required
• Puppeteer bundles Chromium browser
• All processing done with npm packages

High-Quality Output

• 8x resolution seal rendering for print quality (300 DPI)
• SVG to PNG conversion with sharp
• Preserves PDF quality and structure

Development

Commands

npm run dev      # Start development server with Turbopack
npm run build    # Build for production
npm start        # Start production server
npm run lint     # Run ESLint

Project Structure

├── app/
│   ├── api/
│   │   ├── analyze-pdf/    # AI vision analysis endpoint
│   │   ├── apply-seal/     # Seal application endpoint
│   │   └── auth/           # Authentication endpoints
│   ├── login/              # Login page
│   └── page.tsx            # Main application
├── components/
│   ├── Header.tsx          # Top navigation
│   ├── PdfUploader.tsx     # File upload component
│   ├── PdfPreview.tsx      # PDF viewer with seal overlay
│   └── SealEditor.tsx      # Control panel
├── lib/
│   ├── azure-openai.ts     # AI vision integration
│   ├── pdf-to-image.ts     # PDF rendering with Puppeteer
│   ├── pdf-processor.ts    # PDF manipulation
│   └── auth.ts             # JWT authentication
└── types/
    └── index.ts            # TypeScript definitions

Environment Variables

Variable	Description	Required
AZURE_OPENAI_ENDPOINT	Azure OpenAI resource endpoint	Yes
AZURE_OPENAI_API_KEY	Azure OpenAI API key	Yes
AZURE_OPENAI_DEPLOYMENT	Deployment name (use gpt-4.1)	Yes
MASTER_PASSWORD	Application access password	Yes
JWT_SECRET	Secret for JWT token signing	Yes

Security

• Password-protected access with JWT sessions
• 24-hour session expiration
• HTTP-only cookies
• Secure flag enabled in production
• Middleware-based route protection

Contributing

Contributions are welcome! Please ensure all changes:

• Follow existing code style
• Include appropriate error handling
• Work with the generalized AI prompt system
• Maintain PDF coordinate system conventions (origin at bottom-left)

License

This project is private and proprietary.

Support

For issues or questions, please contact the development team.