Recognise text (OCR)
Pulls the text out of a photo, a scan or a PDF made of images. You can have it as a text file or as a searchable PDF — the one where the document looks identical but the text can be searched and copied. The file never leaves your device: the recognition happens right here.
Your file stays on your device: nothing is uploaded
How it works
Recognition is done by Tesseract, Google’s OCR engine, compiled to WebAssembly and served from our own domain: neither the image nor the text passes through a server. PDF pages are drawn at 300 DPI before being read, because below that small type — footnotes, addresses, numbers — becomes unreadable even for a machine. What this tool does not do, and is better known upfront: it does not straighten skewed photos, it does not preserve layout (columns and tables come out as lines of text), it does not read handwriting, and it knows two languages, Italian and English. Above all: an OCR never fails — when it does not understand, it invents, with the same confident face it has when it is right. That is why we always tell you how confident it was, and when that is low we say so plainly.
Your files are never uploaded. Anywhere.
Not to a server of ours, not to a third party, never. When you pick a file it stays in your browser’s memory: it is read, processed and written back on your device, and you save the result yourself. There is no path that takes it elsewhere, because there is no server to send it to.
- Processed locallyFormats are handled by WebAssembly code running in your browser, not by a remote service.
- Nothing to upload, nothing to wait forOther services send your file to a server and then send it back to you. Here the file never leaves: there is no upload to wait for.
- No accountNo sign-up, no email, no daily limits to get around by paying.
- No files to deleteOthers promise to erase your files after an hour. We have nothing to erase: they never reached us.