Secure KYC Auto-Capture

A production-ready React component for auto-detecting and auto-capturing Indian KYC documents (Aadhaar, PAN, Passport, Driving License, etc.) with real-time quality feedback and AI-powered sharpening.

⚡ Fully Offline Solution – All processing happens locally in the browser.
🔐 More Secure by Design – No image data is sent to any server.

✨ Features

• Auto-Capture: Intelligent document detection and automatic capture.
• AI Sharpening: Post-capture sharpening for maximum text clarity.
• Quality Feedback: Real-time stats for Sharpness, Glare, and Document Size.
• Zero Data Transmission: No backend calls. No cloud uploads.
• Privacy First: Local processing using TensorFlow.js (COCO-SSD).
• Responsive: Works on Mobile (1080p ideal) and Desktop.

📦 Installation

npm install secure-kyc-capture

🛠 Usage

import { CameraView } from 'secure-kyc-capture';
import 'secure-kyc-capture/dist/style.css'; // If you use the bundled styles

function App() {
  const handleCapture = (data) => {
    console.log('Captured Image:', data.image);
    console.log('Document Type:', data.docType);
    console.log('Confidence:', data.confidence);
  };

  return (
    <div style={{ width: '100vw', height: '100vh' }}>
      <CameraView onCapture={handleCapture} />
    </div>
  );
}

🔌 Props

| Prop | Type | Description | |------|------|-------------| | onCapture | (data: CaptureData) => void | Callback triggered when a document is captured. |

🧠 CaptureData

{
  image: string;      // Base64 encoded JPEG
  docType: string;    // Detected class (e.g., 'document')
  confidence: number; // AI confidence score
}

🚀 Live Example Application

A complete working example demonstrating integration in a real React + TypeScript app:

🔗 Example App Repository:
https://github.com/PRATHAM-SPS/secure-kyc-capture

This example includes:

• Camera page
• Preview screen
• Retake flow
• Light/Dark mode

👨‍💻 Author

Built and maintained by Pratham Ingawale
GitHub: https://github.com/PRATHAM-SPS

License

MIT