How to get text from an image
- Pick the language of the text, then drop a photo, screenshot or scan into the box, choose a file, or paste a screenshot. You can add several.
- Wait a moment while the text is read. The first time, the reader and the language download (about 5 to 8 MB), which your browser keeps for next time.
- Copy the text or download it as a TXT file. Pick Join into paragraphs for text that wraps across lines, such as a page of a book.
Getting the best results
- Crop to the text. Remove surrounding photos, logos and background so the reader only sees the words. The crop tool helps.
- Straight and sharp. Photograph pages square on, in good light, and hold the camera steady. Tilted, curved or blurry text is the most common cause of errors.
- Big enough. Text should be at least 20 pixels tall in the image. Zoom in on a screenshot before you capture it, or use a higher-resolution photo.
- Dark on light. Black text on white reads best. For light text on a dark background, invert the image first.
What OCR is
Optical character recognition looks at the shapes in an image and works out which letters they are. This tool uses Tesseract, an open-source engine first developed at HP and later maintained by Google and the open-source community. Its newer neural network reader recognises whole lines at once, which is why it copes well with ordinary printed text.
Always read the result through before relying on it. Common slips include the letter O for zero, l for 1 and rn for m.