Key takeaways
- Right-to-left scripts need an OCR engine trained for them — choose the exact language.
- Mixed documents (e.g. Hebrew with English numbers or names) work best with both languages selected.
- Clean, high-contrast scans matter even more for Arabic script, where letters connect.
- Osd-Scan supports Hebrew, Arabic, Persian and Urdu OCR in the browser, with no upload.
Many OCR tools are built for English first. Hebrew, Arabic, Persian and Urdu are written right to left, use different letter shapes and often mix with Latin text and numbers. With the right settings, they can be recognized accurately — and searchable PDFs work just as well as in English.
Why right-to-left OCR is different
- Direction: text runs right to left, but numbers and Latin words inside it run left to right.
- Connected letters (Arabic, Persian, Urdu): letters join and change shape at the start, middle and end of words.
- Small marks: dots and vowel signs distinguish letters, so blur causes more errors.
- Hebrew final forms: letters like ך ם ן ף ץ have special end-of-word shapes.
A dedicated language model handles all of this — which is why choosing the language matters.
How to OCR a Hebrew or Arabic PDF
- Open OCR PDF.
- Add your scanned PDF or photos.
- Choose the language: Hebrew, Arabic, Persian or Urdu. Add English if the document contains English words or names.
- Start recognition. The OCR engine is downloaded once, then runs in your browser.
- Download the searchable PDF.
For new scans, tick Recognize text (OCR) in Scan to PDF and select the language before saving.
Get the best accuracy
- Scan sharp: hold the phone steady and let auto-capture take the photo.
- High contrast: use Auto color or Black & white for printed letters.
- Avoid tiny text: move the phone closer so small print fills more pixels.
- Straight lines: Auto-straighten and flattening help the engine follow each line.
Mixed-language documents
Israeli contracts often mix Hebrew text with English names, emails and numbers. Arabic documents may include French or English. Select both languages — the engine checks each word against both alphabets.
After OCR: search, copy, convert
- Search the PDF in any viewer.
- Copy sentences into emails or messages.
- Convert to Word with PDF to Word or to text with PDF to Text.
- Make it accessible — screen readers can read Hebrew and Arabic text layers aloud.
Privacy for sensitive documents
Many Hebrew and Arabic documents are official: ID copies, court papers, medical results. Osd-Scan's OCR runs on your device and never uploads files.
Right-to-left documents deserve the same convenience as English ones: searchable, copyable and easy to archive.
Frequently asked questions
Can I OCR a document that mixes Hebrew and English?
Yes. Select both Hebrew and English. The engine recognizes both scripts on the same page.
Why is Arabic OCR less accurate than English?
Arabic letters connect and change shape depending on position, and dots are small. A sharp, well-lit scan makes a big difference.
Does the text come out in the right direction?
Yes. Words are stored in logical reading order, so copying and searching work correctly in right-to-left languages.
Can I convert a Hebrew scan to Word?
Yes. Use PDF to Word with OCR enabled and choose Hebrew.