Copy text from an image or screen

How to use

1. Choose an image, paste a copied image, or capture your screen.
2. Drag on the image to select the area to read (the whole image at first).
3. Choose the language of the text (up to 3) and press “Read text”.
4. Edit the result if needed, then copy or save it.

About reading

• Reading uses Tesseract.js (Apache-2.0) and runs on this device. The image and the text are not uploaded or saved.
• The data for each language (about 0.7–3 MB) is loaded from this site the first time you read in that language. No outside service is used.
• The text language is chosen separately from the page language. The first choice is only a suggestion based on the page language.
• With Japanese selected, extra spaces between Japanese characters are removed.
• The tool is meant for printed text. Handwriting, formulas, vertical text, table layouts and text styling are not reproduced. Characters such as 0 and O, or 1 and l, may be misread, and lines may come out of order.

Capturing the screen

• In your browser’s dialog, choose the screen, window or tab to share. Only one frame is captured, and sharing stops right after.
• If you choose the entire screen, the whole screen is passed to this page before you select an area (it is not uploaded). Choosing only the window or tab you need is safest.
• Browsers that cannot capture the screen (most phones, for example) can still use a screenshot: choose it or paste it.

Limits

• Images up to 50 MB and 50 million pixels. The area is scaled to at most 3,200 px on the long side for reading, and small areas are enlarged up to 3 times.
• The area must be at least 16 × 16 px. Up to 3 languages at a time.