Home / Image to text

Image to text

Drop a photo or screenshot – text recognition (OCR) turns it into copyable text. Runs in your browser, no upload.

Drop an image with text here

or tap to choose. Photo, scan or screenshot.

Choose files

Step by step

  1. Choose your document's language under “Language of the text”. There is a separate setting for text that mixes German and English.
  2. Drop a photo, scan or screenshot. Several images are processed one after another.
  3. With the first image, the browser loads the text recognition along with its language data, and the status line shows the progress in percent. As long as you don't change the language, it is reused afterwards.
  4. Each image gets its own card with the recognized text in a text box. You can correct recognition errors right there.
  5. “Copy” puts the recognized text on your clipboard. “Save as .txt” saves the current content of the text box, including your corrections, as a text file.

What you can use it for

  • Reusing screenshots: Take an error message, an address or an order number from a screenshot without typing it out.
  • Digitizing printed documents: Save letters, recipes from cookbooks or notices as text so you can search or edit them later.
  • Quotes from books: Photograph a book page and copy out the passage you need for notes, a presentation or an essay.
  • Foreign-language text: Capture menus, signs or instructions in English, Spanish, French or Italian, then paste the text into a translation program.
  • Scanned PDFs: First convert a PDF without copyable text into images with “PDF to JPG”, then let them be recognized here.

For good results

  • Photograph the document straight on and filling the frame.
  • Even light, no shadows over the text.
  • Choose the right language.

Handwriting is only recognized to a very limited degree. Printed and on-screen text work best.

Frequently asked questions

Are my images uploaded?

No. Text recognition is loaded into your browser once and then runs on your device.

Does it recognize handwriting?

Hardly. It is made for printed text.

Can I process several images at once?

Yes. They are processed one after another, each with its own result.

What does the confidence in the status line mean?

It is the text recognition's own estimate of how certain it is overall. A low value suggests a blurry photo, an unusual typeface or the wrong language. In that case, proofread the text especially carefully.

Are formatting, tables and columns kept?

No. The result is plain text. Line breaks are largely kept, while font sizes, bold text and table structure are lost. For multi-column pages, it helps to photograph each column separately.

Can I recognize text from a PDF?

Not directly. Convert the PDF to images first with “PDF to JPG”. If the PDF already contains selectable text, it's easier to copy it in your PDF viewer.

Which image formats are supported?

Common formats such as JPG, PNG and WebP, as well as iPhone photos in HEIC format. Phone photos taken upright are read the right way round.

What helps with very small print?

Move the camera closer so each letter has enough pixels. For screenshots, zoom into the document or page first and only then take the screenshot.

Why does the result say “No text found”?

The recognition found nothing in the image it could classify as text. Common reasons are heavy blur, very low contrast or text that takes up only a tiny part of the image. In that case, crop the image to the text and try again.