IT
Image to Text (OCR)
Extract text from a photo, screenshot or scanned document — both English and Hindi supported. Processed in your browser.
Advertisement
Input
Choose an image — photo, screenshot, scanned page
a clean, straight, well-lit image works best
Extracted Text
Advertisement
How to use — Image to Text (OCR)
- Upload an image containing text — printed text works best.
- Choose the language (English/Hindi) and click Extract text.
- The language model downloads the first time (~10-15 MB) — after that, processing happens entirely in the browser.
FAQ
Will this read handwriting?
Clean handwriting (block letters) gives decent results, cursive/fast handwriting will have errors. Printed text (books, documents) gets 95%+ accuracy.
Why is the first run slow?
The language model (~10-15 MB) downloads the first time. After that it stays in the browser cache — it'll be fast next time. Processing itself happens on your device.
How do I improve accuracy?
A straight (untilted) photo, good lighting, sharp focus. Use full resolution for screenshots. Crop to just the text if it's small.
Advertisement