Redact PDF

Whether the text is destroyed or merely hidden is precisely what the two modes decide, which is why they are separated rather than blended into one button. In Wipe mode the text is genuinely gone from the pages you marked: those pages are re-rendered as images with the black rectangles burned into the pixels, and the output document is built fresh from those images, so the parsed original page content is never serialised into the file you download. In Cover mode nothing is removed at all. A black rectangle is drawn on top of the page and the words underneath remain fully present — selectable with the mouse, extractable by any text-scraping tool, visible in a hex editor. Cover is for marking up; Wipe is for redacting.

Wipe mode rasterizes each affected page at up to two and a half times its natural size, capped at 2200 pixels wide, so the result stays crisp in print. The black rectangles are painted into that bitmap in canvas pixel space, the bitmap is encoded as a high-quality JPEG, and a new page of the original displayed dimensions is created in a fresh document holding nothing but that image. Pages you never marked are copied across in a single batch, so shared fonts and images are duplicated once rather than per page, and they keep their real text. Document-level structures — outline, attachments, form data, embedded JavaScript and XMP — are deliberately left behind, because any of them can echo the text you just removed.

Whiteout PDF is the white-rectangle equivalent of this tool's Cover mode, and neither of them removes anything. Whiteout draws an opaque white box over a region so the page looks as though the content was never there, which is useful for tidying a form, hiding a letterhead before reusing a template, or blanking a stale figure — but the covered text is still sitting in the content stream, and copying the area or running text extraction will bring it straight back. Wipe mode is the only one of the three that changes what the file actually contains. Pick Whiteout for cosmetic cleanup, Cover for a visible black mark on non-sensitive content, and Wipe whenever the information underneath must not be recoverable.

Only the pages you did not wipe stay searchable. A wiped page is an image, so its text is no longer selectable, searchable or extractable — the same mechanism that makes the redaction effective, and it applies to the whole page rather than only to the black boxes. Pages with no rectangles are copied over as they were and keep their native, fully searchable text. In Cover mode every page stays searchable, including the words behind the black rectangles. If you need the remaining wording of a wiped page back in searchable form, run the redacted file through OCR PDF, which recognises the visible text and lays a text layer over the image.

A wiped document carries almost no hidden data forward, by design, while a covered one carries all of it. Wipe keeps only the descriptive basics — title, author, subject, keywords, creator and the original creation date, with the modification date set to now — while bookmarks, attachments, form data, embedded JavaScript and the XMP metadata packet are dropped rather than copied, since each can repeat the exact text you just removed. Cover mode changes nothing except adding rectangles, so all of those structures survive intact along with the covered text. Whichever mode you use, check the result before sending it: open it, try selecting the redacted area, and inspect the fields with Edit PDF metadata.

Redact PDF is free, has no limits and adds no watermark, and no part of the document is uploaded — which for this tool is the entire point. Pages are rendered with pdf.js, rectangles are drawn against the canvas in front of you, and the rebuilt file is assembled with pdf-lib, all inside your browser tab; the sensitive original, the intermediate page images and the redacted output all stay on your machine. Nothing is queued on a server, so there is no copy anywhere to be retained. Encrypted files have to be unlocked first with Remove PDF password, since page content has to be parsed before it can be rasterized. Local processing does cost time on a long document, because every affected page is rendered at high resolution before the file can be rebuilt around it.