Text from image (OCR)
Recognises the text in a photo, a screenshot or a scanned PDF and gives it back as text to copy or download. Russian and English. Recognition runs in the browser tab: the picture is not uploaded; the language models (about 3 MB each) download once on first use.
- In
- JPG · PNG · WEBP · PDF
- Out
- text
Files are processed in this tab and are not uploaded anywhere.
Related
Questions
How do I get the best result?
A sharp, straight, well-lit picture of printed text. Photograph the page from above, fill the frame with it, avoid shadows. Handwriting is not recognised reliably.
Does it work on PDFs?
Yes — every page is drawn at 300 dpi and read. If the PDF already has a text layer (you can select the words), PDF → text gets it faster and exactly.
Why is the first run slow?
The recognition engine and the language models (about 15 MB together) download once and stay in the browser’s cache; next time it starts at once.