For the best result
- Photograph the page straight on, in good light with no shadow.
- Pick the right language, and Arabic and English for mixed documents.
- Printed text reads well. Handwriting gives poor results.
In our test on a clear image of lines from an Arabic invoice, the tool read every word and number correctly, and the same for English text. Tilted or dim phone photos come out with more mistakes, so check important names and numbers by eye.
Scanned PDFs
You can choose a scanned PDF and its pages are read one after another, up to 30 pages. If the PDF holds real text, PDF to text is faster and more accurate.
Common mistakes
- Choosing the wrong language: Arabic read with English only gives nonsense. Pick Arabic and English for mixed text.
- Using a tilted, blurry or dark photo: retake it straight on, close up, in good light.
- Using OCR on a PDF that already contains text: PDF to text is faster and exact.
Your photos never leave your device
Everything happens inside your browser. Your images are not uploaded to a server and we keep no copy, so you can use the tool for ID photos and private pictures. There is no sign-up and no daily limit.
Frequently asked questions
Does it work offline?
It needs the internet the first time to download the language data. Your image is never sent anywhere.
Why are some words wrong?
Usually a blurry or tilted photo, or very small print. Take the photo again from closer and in better light.
Can it read handwriting?
It is built for printed text and reads handwriting poorly.