Perso-Arabic (Nastaliq) script
Urdu OCR اردو
Extract text from scanned Urdu documents and images. Urdu is written right to left, usually in the flowing Nastaliq style, and the result comes out as Unicode text you can edit or give to an AI assistant.
● Runs in your browser. Your files are never uploaded..pdf.jpg.png.docx.xlsx.xls.csv.txtUrdu and English are already selected. Add or remove languages in the box above.
Good for
- Urdu newspapers and books
- Old property and court documents
- Religious and literary texts
- Letters and notices
Tips for better Urdu results
- Nastaliq is one of the hardest scripts for OCR: expect more errors than with Hindi or English, especially in dense or decorative text. Naskh-style printing converts better.
- Use the highest-resolution scan you can get.
- Urdu and English are selected together. Remove English for purely Urdu pages.
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 Urdu 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 Urdu 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 Urdu documents uploaded anywhere?
No. OCR runs inside your browser. The Urdu language data is downloaded to your browser once, but your files never leave your device.
How accurate is Urdu OCR?
Clean, printed Urdu 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.