Image to text: copy text from a photo, scan or PDF

Read the text in a photo, a scanned page or a PDF. Copy it, save it as a .txt file, or make a searchable PDF. The reading happens on your device.

Add a photo, scan or PDF

We read the text on it, here on your device. Then copy it, save it as a .txt file or as a searchable PDF.

JPG, PNG, HEIC, or a PDF (scanned pages are fine).

Files stay on this device. Nothing is uploaded.

Text reader

Reading text needs a reader: 27.5 MB the first time, English included. It is downloaded once from this site and kept on your device; nothing you scan is sent anywhere.

Language

Good to know

  • Works best on clear printed text. Handwriting is often wrong.
  • The words come out in reading order, but tables and columns are not kept.
  • Always check numbers before you use them.

How it works

First the reader finds where the lines of text are on the page, then it reads each line letter by letter. Both steps are small neural networks (PaddleOCR PP-OCRv5, an open-source project) that run in your browser using ONNX Runtime.

Tips for better results

  • Use the document scanner first: a straight, clean page reads much better than a tilted photo.
  • Good light and a sharp photo matter more than anything else.
  • Pick the language of the text before reading.

Honest limits

  • Mistakes happen, especially with small print, stamps and handwriting. Check names and numbers.
  • Tables and columns are read line by line; the layout is not kept.
  • The first use downloads the reader (shown with its size). After that it is instant and works offline.

Questions people ask

Is my document uploaded to read the text?

No. The text reader is downloaded once from this site to your device, and the reading happens there. Your photo or PDF is never sent anywhere.

Why does it need a download of about 27 MB?

Reading text needs a small neural network (the reader). It is downloaded only when you tap the button, shown with its size first, and kept on your device, so the next time it works at once, even offline.

Can it read Hindi?

Yes, choose "Hindi and Marathi (Devanagari)". It adds about 8 MB to the reader. Hindi text can be copied and saved as .txt; the searchable PDF currently works for English only.

Can it read handwriting?

Not reliably. It is made for printed text such as certificates, admit cards, notices and forms.

Can I turn a scanned PDF into Word with the same layout?

No. You get the words in reading order, not the layout, tables or fonts. Copy the text into your document and format it there.

Related tools