Birbal AI

Free tool · Arabic OCR

Free Arabic OCR image and PDF to text, in your browser.

Turn a photo, scan or PDF with Arabic or English text into text you can copy and edit. The OCR engine runs on your own device, so your files are never uploaded, and the words it is less sure of are highlighted for you to check.

  • Arabic, English or both on one page
  • JPG, PNG, WebP and PDF, never uploaded
  • Unsure words highlighted for checking

Arabic OCR

Extract Arabic text from an image or PDF

Choose a file or try the sample page. The first run downloads the OCR engine and the Arabic and English data, about 9 MB, from this site; after that it starts straight away.

JPG, PNG, WebP or PDF, up to 25 MB a file and 20 pages a PDF.

Runs in your browser: your files are not uploaded, and nothing you read here is sent to us.

The text appears here, ready to copy or download, with any words to check highlighted below it.

Authorized vendor to

  • Ministry of Finance
  • Ministry of Interior
  • DGE Abu Dhabi
  • Etihad Aviation
  • EDGE Group

Delivered for

  • Hindustan Unilever
  • Nokia
  • Linde
  • USA Today
  • JSL

Procurement-ready

  • Emirati national company
  • National ICV certificate
  • Abu Dhabi Government Procurement Gate
  • UAE Ministry of Finance, federal procurement
  • Registered ADNOC vendor

How it works

Open-source OCR, running on your device

Tesseract in your browser

The tool runs Tesseract, the open-source OCR engine, with its Arabic and English language models. It loads from birbalai.com and reads the page on your device.

PDFs page by page

Each PDF page is drawn as an image and read in turn, up to 20 pages at a time. Photos are straightened by the browser and resized for reading.

A confidence score for every word

Tesseract scores each word it reads. Words below 70 out of 100 are highlighted, so you know where to look before you use the text.

Clean text

Getting good Arabic text out of a scan

Scan at 300 dpi

Tesseract reads best at 300 dpi or more. Small or blurry letters lose their dots, so scan at that setting, or take a sharp photo with the page filling the frame.

Keep the page flat and evenly lit

Photograph straight on, without shadows, glare or a curved book spine. Crop away the desk around the page.

Pick the right language

Choose Arabic only for pages with no English, and Arabic and English for mixed letters, invoices and forms.

Know the limits

The tool reads printed text. Handwriting, heavy diacritics, stamps and lines through the text need checking, and tables come out as plain lines of text.

Book a demo

Turn your whole archive into searchable text

Wathiqa reads Arabic and English scans as they arrive, scores its confidence on every word and keeps the documents searchable, in a UAE cloud region or on your own servers. Book a 30-minute video demo and bring a few of your own scans.

  • A live demo on your own files
  • Arabic and English
  • In the UAE or on your servers

4.9/5 average rating · 98% client satisfaction · 150+ clients served

Book a Wathiqa demo

We only use your details to reply to you. Arabic or English, as you prefer.

FAQ

Arabic OCR, answered

Is this Arabic OCR free?

Yes. There is no sign-up and no daily limit. PDFs are read up to 20 pages at a time, and you can run as many files as you like.

Are my files uploaded?

No. The OCR engine and the language data download to your browser from birbalai.com, and the page is read on your device. Your files and the text stay there, and are gone when you close the page.

Which files can I use?

JPG, PNG, WebP, BMP and GIF images, and PDFs, up to 25 MB a file; the first 20 pages of a PDF are read. Phone photos work when the page is sharp and fills the frame. If an iPhone photo will not open, save it as JPG first.

Can it read handwritten Arabic?

Not reliably. Tesseract is made for printed text. For handwriting, our Arabic OCR comparison, linked below, shows which engines list Arabic handwriting.

Why are some words highlighted?

Tesseract gives every word a confidence score from 0 to 100. Words below 70 are highlighted so you can check them against the page. Small print, missing dots and diacritics, stamps and lines through the text are common causes.

Does it keep the layout or tables?

No. It returns plain text in reading order, and a table comes out as lines of text. To keep the scan with its text and search both, use a document management system such as Wathiqa.

How do we OCR thousands of documents?

Use a system built for it. Wathiqa reads Arabic and English scans as they come in, scores its confidence on every word, gives your team a review screen for the words it is unsure of, and runs in a UAE cloud region or on your own servers.

Make every Arabic scan searchable

Book a live Wathiqa demo with your own documents.

Book a Wathiqa demo