Arabic script

Arabic OCR العربية

Extract text from scanned Arabic documents and images. Arabic is written right to left and letters change shape with their position in a word; the Arabic OCR model handles the joined forms and returns Unicode text.

● Runs in your browser. Your files are never uploaded.

Loading your limits…

Drop files here, or tap to choose PDF, scans, photos, Word, Excel, CSV. You can also paste a screenshot.

.pdf.jpg.png.docx.xlsx.xls.csv.txt

Good for

Contracts and official letters
Arabic books and newspapers
Certificates and official documents
Invoices from Gulf countries

Tips for better Arabic results

Printed Naskh text converts best. Decorative or calligraphic styles are much harder.
Short vowel marks (harakat) are often dropped; that is normal for OCR.
Arabic and English are selected together for bilingual documents.

How it works

Drop a scanned PDF, photo or screenshot into the box above, or paste an image.
Pages that already contain text are read directly. Scanned pages go through Arabic OCR in your browser.
Download the result as a Markdown (.md) file, or copy it and paste it into ChatGPT, Claude, Word or Google Docs.

Questions

Is this Arabic OCR free?

Yes. Without an account you can convert 10 scanned pages a day, and a free account gives you 30. Pro removes the limit.

Are my Arabic documents uploaded anywhere?

No. OCR runs inside your browser. The Arabic language data is downloaded to your browser once, but your files never leave your device.

How accurate is Arabic OCR?

Clean, printed Arabic text at 300 dpi converts well. Handwriting, faint photocopies, stamps and decorative fonts reduce accuracy, so check names, numbers and dates against the original.

Other languages

Read our guides on getting better results from documents in AI tools.