Image to text (OCR)
Turn a photo of a document, a screenshot or a scan into editable text. Recognition runs on your device and supports 20 languages.
Runs in your browser. The open-source OCR engine (about 5 MB) and language data (1 to 4 MB each) download once from jsDelivr. Your images never leave your device.
Frequently asked questions
How accurate is it?
Clear, well-lit text in a straight line is usually recognised almost perfectly. Handwriting, very small text and photos at an angle are harder. Pick the correct language for the best result.
Can it read Arabic and other languages?
Yes. Choose up to four languages at once, including Arabic, Urdu, Persian, Hindi, Chinese, Japanese and Korean.
Are my images uploaded?
No. Only the OCR engine and language files are downloaded. Your images are read on your device.