Image to text (OCR)

Turn a photo, screenshot or scanned PDF into text you can copy and edit. The reading is done by your own browser, so your documents are never uploaded.

How to extract text from an image

  1. Choose or drop one or more pictures, a scanned PDF, or paste a screenshot.
  2. Pick the language the text is written in, then click Extract Text.
  3. Fix anything that was misread, then click Copy Text or download it as a TXT file.

What is OCR?

OCR stands for optical character recognition: software that looks at a picture of words and works out which letters they are. It is how you get editable text out of a photo of a page, a screenshot, a receipt or a PDF that was made with a scanner.

Your files are never uploaded

Most online OCR sites send your picture to their servers. This one runs the open-source Tesseract engine inside your web browser, so the reading happens on your own computer or phone. That makes it safe for IDs, bank letters, contracts and anything else you would not want to hand to a stranger. There is no sign-up, no page limit and no watermark.

The first time you use it your browser downloads the engine (about 4 MB) and the data for your language (1 to 3 MB each). After that they are kept on your device and it starts straight away.

How to get the best result

  • Sharp and straight. Hold the camera square to the page. Tilted, curved or blurry text is the main cause of mistakes.
  • Even light. Avoid shadows across the page and the glare of a flash on glossy paper.
  • Big enough letters. Crop to the part you need rather than photographing a whole desk. Tiny text in a large photo is hard to read.
  • Choose the right language. Accents and other alphabets are only recognized when their language is selected. For mixed documents, add a second language.

What it can and cannot read

Printed text in books, letters, screenshots, slides, signs and scanned documents usually comes out very well. It is honest to say where it struggles:

  • Handwriting is mostly not recognized. The engine is trained on printed type.
  • Tables and columns are read as plain lines of text, so the layout is not kept.
  • Decorative fonts, low-resolution photos and text over busy backgrounds give more errors. Always proofread anything important, especially numbers.

How fast is it?

A screenshot or a single page takes a few seconds on a computer and a little longer on a phone. A scanned PDF is read page by page, at roughly 3 to 10 seconds per page, so a long document can take a few minutes. Keep the tab open while it works; you can press Cancel and keep the pages already read.

If your PDF already has selectable text, use PDF to Text instead: it is instant and exact. To turn pages into pictures use PDF to Image, and count what you extracted with the word counter.

Frequently asked questions

Is this OCR tool free?

Yes. It is free with no sign-up, no daily limit and no watermark.

Are my pictures or PDFs uploaded?

No. The text is read by your own browser and your files never leave your device.

Which languages are supported?

English, Spanish, French, German, Italian, Portuguese, Dutch, Swedish, Polish, Turkish, Russian, Ukrainian, Arabic, Hindi, Indonesian, Vietnamese, Chinese (Simplified and Traditional), Japanese and Korean.

Can it read handwriting?

Not reliably. It is built for printed text. Very neat block handwriting sometimes works, but joined-up writing usually does not.

Can I extract text from a scanned PDF?

Yes. Add the PDF and each page is turned into a picture and read in turn. Long documents take a few minutes because the work is done by your device.

Why are there mistakes in the text?

OCR depends on the picture. Blurry, tilted, dark or very small text causes errors. Retake the photo straight on in good light, crop to the text and make sure the right language is selected.