Private · on device. Runs in your browser: your file is not uploaded. No sign-up.
A scanner saves a picture of each page. You can read it, but your computer cannot: search finds nothing, copying gives nothing, and a screen reader stays silent. OCR (optical character recognition) reads the letters in the picture and stores them alongside it.
No. Every page keeps its original image. The recognised words are added as an invisible layer exactly over the printed words, so the PDF looks the same and can now be searched, selected and copied.
They are kept exactly as they are, which is faster and more accurate than reading them again. Turn on "Read every page" if a page has broken or wrong text that you want replaced.
No. Pages are rendered and read in your browser, and the new PDF is built there too. Nothing is sent to a server.
Roughly two to six seconds per page on a modern laptop, longer on a phone. The first run of a language also downloads its model once (1 to 10 MB), which the browser then keeps.
