The most consequential mistake in this whole category is drawing a black rectangle over a name and believing it has been removed. It has not. The text is still in the file, sitting underneath a shape, and anyone who selects the area and pastes it somewhere gets the original wording back. This has produced real, published, professionally embarrassing failures, repeatedly — and it is exactly the mistake this tool exists to make impossible.
So the first thing to be clear about is the difference between two operations that look identical on screen. Covering places an opaque box over an area: the content vanishes visually, but the characters stay in the file, still selectable, still found by search, still recoverable in seconds. Removal deletes the characters themselves, so there is nothing left underneath. They are not the same, and treating a cover as a removal is where the damage happens. The cover tool here is labelled as exactly that — covering — and this tool is the other one.
What you do here is mark: type the name, number or phrase that must go, and it is swept across every page — because a name in a witness statement shows up in the first line, the signature block and the address at the bottom, and removing one of three achieves nothing. Or draw a box over an area, and the marks outline on the page in a way that is visibly distinct from a committed cover. The panel lists what the sweep found — every instance, every page — before anything happens.
Then the run asks once, plainly: these are deleted from the new file itself, not covered, and there is no getting them back from it. It produces a fresh file — your original is never touched — and a receipt of exactly what was removed. The fresh file carries none of the original’s hidden details either: no metadata, no notes, no attachments.
And when it cannot do the job honestly, it refuses. If a marked word cannot be traced to its exact bytes, if the document has optional layers that a fresh copy could expose, or if an image under a mark cannot be rewritten, the run produces nothing and names the reason — it never hands you a file that merely looks redacted. Scanned pages carry no text at all, only an image, so mark the area there; where the pixels can be proven gone they are erased, and where they cannot, the run says so.