Amazon Textract

Arabic print since November 2026; handwriting English-only, no Gulf regions

Visit website →

About

Textract is the OCR and document-analysis service of AWS, extracting text, tables, forms, signatures and answers to natural-language queries from scans and PDFs. Until recently it did not read Arabic at all. An announcement on 2 November 2026 expanded the text-detection model to more than forty languages including Arabic, and the FAQ now says the same, but the supported-languages page still places Arabic in its third category, covering tables, layout, signatures and asynchronous text detection; Queries, expense analysis and identity analysis stay English-only, Forms covers six Latin-script languages, and handwriting "is only supported in English". Identity analysis reads US government documents only. Pricing in US regions in November 2026 is US$1.50 per thousand pages for text detection, US$15 for tables and US$50 for forms, with a three-month free tier of a thousand pages a month. Textract is available in eighteen regions; neither Bahrain nor the UAE is among them, so Saudi pages are processed in Europe or the US.

AI-assisted draft, human-reviewed before publishing — see how we choose & review tools.

Why we picked it

Most Arabic OCR comparisons still say Textract does not support Arabic, and until this month they were right. We list it on the day the status changed so AWS teams in the Kingdom know the new scope and its limits: text and tables only, English handwriting, no Gulf region.

Featured in our guides