How to extract text from an image
- Select Choose files, or drop a photo, scan or screenshot on the box.
- Wait while the tool reads the picture. The first image takes a few seconds longer because it loads the text engine.
- Check the text shown in the result. Select Copy to put it on your clipboard, or Download to save it as a .txt file.
What OCR does
OCR stands for optical character recognition. The tool looks at the shapes in your picture, finds lines of text and turns each shape into a letter. The result is plain text that you can search, edit, translate or paste into another document. It is useful for screenshots of chats, photos of printed pages, scanned receipts, slides and signs.
How it runs in your browser
The recognition engine is Tesseract, compiled to WebAssembly. Its worker, its core and its English language data are stored on this website, so the page does not call any other server while it works. Your picture is decoded on your device, handed to the engine and discarded afterward. Only the text you copy leaves the page, and only because you copied it.
Reading a large photo can take several seconds on a phone. The progress bar shows the stage the tool is in.
Tips for better text
- Crop the picture to the text with the crop tool so the engine does not waste effort on the background.
- Make small text easier to read by enlarging the image with the image resizer.
- Remove colour and noise first with grayscale when the page has a coloured background.
- Straighten sideways photos with rotate image before reading them.
- Need the pages as a document instead of text? Combine them with images to PDF.
Limits
OCR can mistake similar shapes, such as the letter O and the digit 0, so always proofread numbers and names. Tables are read line by line, so columns may run together. Handwriting is not supported.
Everything happens on your device. Your images and the text found in them are never uploaded.