Prodact

Prodact blacks out parts of a PDF and gives you back a new file where the covered words are genuinely gone. Most tools draw a black rectangle over the page. The text underneath is still in the file: it can be selected, copied, searched, or read by anything that opens the document. Prodact does not do that. Each page is turned into an image, the black rectangles are burned into the pixels, and the new file is built from nothing but those pixels. Nothing is copied over from the original — no text, no fonts, no metadata, no earlier versions of the page. WHAT IT DOES • Draw a box over anything you want covered, on any page. • Search the document for a name, a number or a phrase, and review every match before anything is applied. Each match tells you which page it is on and whether it came from the document's own text or from reading the page as a picture. • Redact one thing and be offered the rest: other places with the same text, the same shape (a phone number, a reference code), or the same position on the line. • Export, and read what was actually done: how many pages were rebuilt, how many rectangles were burned in, how many words were checked, and whether the finished file still contains anything you searched for. EVERYTHING HAPPENS ON YOUR PHONE Prodact does not ask for the internet permission, so Android will not let it open a network connection. There are no accounts, no analytics, no advertising, and nothing to sign up for. Text recognition runs on your device from a model built into the app. Android's backup is switched off for this app, so your list of documents does not travel to the cloud or to a new phone. The screen showing your pages is marked secure, so page content stays out of screenshots and out of the thumbnail Android takes when you switch apps. IT TELLS YOU WHAT IT CANNOT DO After an export, Prodact reads its own output back and tells you what it found — including when it found a term you had searched for still readable in the finished file. It distinguishes between "we searched and found nothing" and "nothing could be searched", because those are the same zero and different facts. A black box hides the words and not their length or their position, and a scanned page is read by recognising shapes, which is not the same as reading text. Prodact says so, on the screen, rather than in a document nobody reads.