OCR Lens

An OCR Engine Comparison platform that extracts text from images in the browser in real time.

• Tesseract.js — Runs entirely on-device (privacy-friendly, no server needed)
• OCR.space — Cloud-based, uses included API key
• Google Vision — Enterprise-grade, requires your own API key

Live Demo

https://vermashaurya.github.io/ocr-lens/

Features

• Image upload via file picker
• Mobile camera capture (rear camera opens directly)
• Live camera with "take photo" flow
• Three OCR engine options to compare
• Confidence score display

OCR Engine Comparison

Engine	Type	Privacy	Accuracy	Setup
Tesseract.js	On-Device	✅ 100% private	Moderate	None
OCR.space	Cloud	❌ Sends to server	Good	Included key
Google Vision	Cloud	❌ Sends to server	Excellent	Requires API key

Run locally

Because camera access is usually restricted on plain file:// pages, serve the folder through a local web server.

Option 1

python3 -m http.server 8000

Then open http://localhost:8000.

Deploy

This app is static. You can deploy the folder directly to services such as Netlify, Vercel, GitHub Pages, or any web server.

API Keys

• OCR.space: Pre-configured with a shared API key (rate limited)
• Google Vision: Enter your own API key in the input field. Get one free at console.cloud.google.com (1,000 requests/month free, requires billing enabled)

Notes

• OCR accuracy improves with sharp, high-contrast images
• The app uses English recognition (eng)
• Tesseract.js is loaded from CDN, so the site needs internet access

Tech Stack

• Tesseract.js — Client-side OCR (CDN)
• OCR.space API — Cloud OCR
• Google Vision API — Cloud OCR
• HTML5 — Semantic markup
• CSS3 — Custom properties, grid layout, responsive design
• JavaScript — Vanilla JS, no frameworks

License

This project is licensed under the Apache License 2.0.

Copyright © 2026



Feel free to take inspiration.
Happy Coding!