Redact PDF

Redact PDF blacks out names, numbers and anything else for good. Drawing a black box over text is not enough — the text stays in the file under it, and anyone can copy it out; many "redacted" court papers have leaked that way. Here, each page with a redaction becomes a picture of itself with the marked areas painted black before the picture is taken, so what was under them is not in the new file in any form. Then the file is opened again, and those pages must have no text left at all.

1. Choose a PDF, or drop it on the page. The Redact tool opens on the first page.
2. To find text, type the words or numbers to remove, one per line, and tick any patterns — email addresses, phone numbers, card numbers, PAN or Aadhaar numbers. Press Find, read the matches, and press Mark all.
3. To mark by hand, press Draw boxes and drag across what should go. Page Up and Page Down change page. Marks show see-through so you can check what each covers; a mark can be moved, resized or deleted.
4. Press Redact and download. The marked pages are replaced, the file is checked, and offered.

• A redacted page becomes a picture at 200 dpi: its remaining text can no longer be selected, searched or copied, and its links stop working. Run OCR afterwards to make it searchable again. Pages without marks are copied as they were.
• Find reads the text a PDF stores as text. Words inside scans and photos are pictures, so mark those by hand. Positions within a line are estimated, so each found box is made a little wider than the word; check every box before downloading — an uncovered letter stays.
• What is under a mark is removed from that page. The same name elsewhere — on another page, in the document properties, or in a file name — is not: search the whole document, and use the Properties tool to clear the title and author.
• Files up to 300 MB. Bookmarks and attachments of the original are not carried over, and a digital signature in the original is no longer valid.
• Your file never leaves this browser tab: pages are drawn and redacted here, and nothing is uploaded or stored. The words you search for are used in the tab only. The page downloads its own code, and pdf.js's fonts and image decoders when a page needs them, from toolsdev.in; no request carries your file. Closing the tab ends the session.

The new file is opened again before it is offered, and every redacted page must give no text at all, the way copying or a text extractor would read it. A file that fails is not offered. You can check yourself: open it, select all on a redacted page, and copy — nothing comes.

The result is as safe: nothing under a mark remains. Acrobat rewrites the page and keeps the rest of its text selectable; here the whole page becomes a picture, which is simpler and leaves nothing to chance, at the cost of that page's selectable text.

Before downloading, yes: marks are in the undo history and can be deleted. After, no — the downloaded file no longer contains what was under them. Keep your original.

Email addresses; phone numbers of ten or more digits, with or without a country code; card numbers of 13 to 19 digits; Indian PAN numbers; and Aadhaar numbers written in groups of four.