Image to Text (OCR)

Free OCR online: extract text from images, screenshots, photos and scanned PDFs in 24 languages. It runs in your browser and nothing is uploaded.

You can also drag files here or paste a screenshot with Ctrl+V (⌘+V on a Mac). The text is recognised by your browser. Files are not sent anywhere: no upload, no storage.
Each language is downloaded once (1 to 3 MB) and then stays saved in your browser. Add English as a second language when the text mixes in words in the Latin alphabet.

What it does

This tool uses OCR (optical character recognition) to turn the text in a picture into text you can copy, edit and search. It works with screenshots, photos of documents, scanned pages and PDFs that are only images. The recognition is done by Tesseract, an open-source OCR engine, running in your own browser, so the files never leave your device. Small text is enlarged automatically before it is read, which noticeably reduces errors.

How to use it

  1. Choose one or more images or a PDF, drag them here, or paste a screenshot.
  2. Check the language of the text. Reading starts on its own; if you change the language, press Extract text again.
  3. Review the result, fix anything that was misread, then copy it or download it as a .txt file.

Frequently asked questions

Are my images uploaded?
No. The recognition engine and the language data are downloaded to your browser, and the images and PDFs are processed there. Nothing is sent to a server, so it is safe to use with private documents.

How accurate is it?
On clear printed text, languages written in the Latin alphabet usually get more than 99% of characters right, and other scripts typically between 93% and 99%. Accuracy drops with blurry photos, low contrast, curved pages, decorative fonts and handwriting. For the best result, use a sharp, well-lit and straight image, and choose the right language.

Which languages are supported?
24 languages: English, Portuguese, Spanish, French, German, Italian, Swedish, Afrikaans, Turkish, Indonesian, Malay, Filipino, Vietnamese, Russian, Arabic, Urdu, Hindi, Tamil, Thai, Khmer, Japanese, Korean, and Simplified and Traditional Chinese. You can combine two languages when a text mixes them.

Can it read handwriting?
Only poorly. The engine is trained on printed text: neat block capitals sometimes work, but joined-up handwriting usually comes out wrong.

Does it work with scanned PDFs?
Yes. Each page is rendered at high resolution and read in turn, and you can mark where each page starts. If the PDF already contains text you can select, the PDF to Text tool is faster and exact.

Wikipedia — Optical character recognition
Wikipedia — Tesseract (software)