Free tool · Arabic OCR
Free Arabic OCR: image to text
Upload a photo, scan or screenshot of Arabic text and get editable text back. Arabic and English can share the page: each line stays in its own language, in reading order.
Your image is sent to our contracted AI model provider to read the text. We don't save the image or the text. To apply the daily limit, we keep only the time and a hashed form of your network address.
For the best results
- Photograph the page flat, straight on, in good light.
- Crop to the text you need. Less clutter means fewer mistakes.
- One page per image. For a PDF, take a screenshot of each page.
- Printed text works best. Handwriting is read too, but expect more mistakes.
From a team that builds Arabic document reading into business systems
We build Arabic document reading into the systems companies run on: invoices, contracts, official letters and forms. There, a misread digit can't be allowed through, so results are checked against the company's own records and a person reviews anything uncertain.
This free tool is the simple version: one image in, plain text out. It is set up to read Arabic right to left, keep Arabic and English in their original languages, copy numbers and dates as written, and write [unreadable] where it can't make out the text instead of guessing. It can still make mistakes, so check anything that matters.
Arabic OCR — common questions
Is this Arabic OCR tool free?
Yes. You can convert up to 10 images a day for free, with no sign-up. The tool also has an overall daily limit, so on a busy day it may ask you to try again later. If you need documents read in volume, we build that into your own systems.
Do you store my images?
No. Your image is sent to our contracted AI model provider to read the text, and we don't save the image or the text that comes back. To apply the daily limit, we keep a usage record: the time, and a hashed form of your network address rather than the address itself. Usage records are deleted after 30 days.
What do I get back?
Plain text you can copy and edit. The tool is set up to return only the text: no formatting, no translation and no commentary. If the image has no text in it, it replies [no text found].
Does it work with Arabic and English on the same page?
Yes. Invoices, IDs and official letters often mix both, so the tool is set up to keep each line in its original language, in reading order, and to copy numbers and dates as written rather than reformat them. It does not translate.
Can it read handwritten Arabic?
It will try, but expect more mistakes than with printed text, especially with rushed or joined-up writing. Where it can't make out the text, it is set up to write [unreadable] rather than guess.
Can I convert a PDF?
Not directly. The tool takes JPG, PNG or WebP images, so upload a screenshot or photo of each page. Large phone photos are resized in your browser before upload. Reading multi-page PDFs in volume is the kind of system we build for companies.
How accurate is it?
It depends on the image, and we don't publish an accuracy figure. Clean printed text in a sharp, well-lit photo gives the best results; blur, glare, low light, skewed photos and small print all cause mistakes. Check any name, number or date that matters. For business use where errors have a cost, we add checks against your own records and a person who reviews anything uncertain.
Need this at volume, inside your own systems?
We build Arabic document reading into the systems you already run, with checks against your records and a person reviewing anything uncertain.
Talk to usYou decide. We answer for it.