Text from a photo
Pick or take a photo of printed Arabic or English text, and get text you can edit, copy or save. The reading happens on your device.
Drop a photo here, paste one, or:
Runs on your device: the photo is never uploaded or saved. The first time, the page downloads the text reader (about 9 MB) from this site.
Getting a good read
The reader is trained on printed text. Handwriting, decorative calligraphy and very small print usually come out wrong.
Hold the phone parallel to the page, fill the frame with the text and keep it in focus. A tilted page reads much worse: the document scanner can straighten it first.
Choose only Arabic or only English when the text is in one language. Mixing both helps with mixed text but can turn some English words into Arabic letters, or the other way round.
Arabic letters join up and change shape with their position, and dots decide the letter: ب ت ث differ only in dots. A speck of dust can change a word.
Sources: Tesseract OCR, an open-source text recognition engine (Apache License 2.0), run in the browser with tesseract.js (Apache 2.0) and its Arabic and English trained data; The Unicode Standard, chapter 9 (Middle East-I), on Arabic cursive joining and the isolated, initial, medial and final forms of letters. Licences: tesseract.js, Tesseract, notices. Results depend on the photo: always check them.