Do I need an account?
No. You get 10 scans per day without signing in. Sign in to earn referral bonuses, or upgrade for history and more scans.
Turn any image into clean, editable text.
Built for the way you actually read: a WhatsApp screenshot in Arabic, a Hebrew invoice, an English contract, sometimes all three on one line. Recognition runs on your device.
Bidirectional by default
Every line is shown with plaintext bidi isolation, so a Hebrew sentence with a price in dollars or an Arabic address with a postcode reads the way it was written. Switch to forced RTL or LTR when you need to.
Numbers and currencies stay intact
₪, $, € and Arabic-Indic digits are kept as recognised rather than normalised away, so prices and reference numbers survive the trip.
Frequently asked questions
Are my images uploaded to a server?
No. Text recognition runs in your browser with Tesseract. Only the scan count, and the text itself if you are on a paid plan and choose to save it, reach our server.
How accurate is the OCR?
Clear, high-contrast printed text usually comes out well. Low light, blur, decorative fonts, small text and handwriting reduce accuracy. Always review the result before relying on it.
Which languages are supported?
Arabic, Hebrew and English are recognised together in one pass, including lines that mix them.
Open the SnapText tool
OCR runs locally in your browser. Enable JavaScript to select or paste an image and review the extracted text.