How to permanently redact a PDF
- Open your PDF. Drop the file onto the page or tap Choose a file. Password-protected PDFs ask for the password right there.
- Mark what has to go. Drag a box over anything sensitive, or type the text — a name, an SSN, an account number — hit Find, step through the matches, and mark them all at once.
- Adjust the marks. Click a mark to move it or resize it by its handles; zoom to 200% for fine placement. Cmd/Ctrl-Z undoes any step. Tick “draw on every page” for headers and footers.
- Redact & download. Pick an output resolution and download. Every page is re-rendered as a flat image with the boxes burned in solid black — open the result and try to select text under a box. There is none.
What it does
- Find & mark: type a name or number, review every occurrence with a prev/next stepper and per-match checkboxes, then box them all at once
- Draw with mouse or finger, then move or resize any mark with drag handles — full undo/redo (Cmd/Ctrl-Z)
- True removal — pages are flattened to images, so there is no text under the boxes to copy out
- Zoom to 200% for precise marks; a thumbnail rail shows every page with its redaction count
- Choose output sharpness: 150, 200 or 300 dpi
- Opens password-protected PDFs and plain images too; everything runs on your device
Why overlay “redaction” keeps failing people
Every year or two the same story runs again: a court filing, a government report or a company disclosure is published with black rectangles over the sensitive parts — and within hours someone has copy-pasted the text straight out from under them. The Manafort filings, redacted school-shooting reports, sealed deposition excerpts: same mistake every time. The rectangle was a drawing sitting on top of the text. The text was still in the file.
That failure is built into how most PDF editors work. A PDF page is a stack of objects — text runs, images, shapes. Drawing a black box adds one more shape to the stack. Unless the editor also deletes the objects underneath (most free ones don't), the “redacted” content ships inside the file, one select-all away from being read.
What this tool does instead
This tool takes the only approach that cannot leak: it flattens. Each page is rendered to pixels at about 200 dpi, your marks are burned in as solid black, and a brand-new PDF is built from those flat images. The original text objects, fonts, hidden layers and metadata are simply not carried over. There is nothing under the boxes because there is no “under” anymore.
The tradeoffs are honest ones. The output is not searchable, because there is no text layer left — that absence is the guarantee. The file usually gets bigger, because image pages compress less well than text. And the redaction is exactly as good as your marks: the tool destroys what you cover, so check the preview before you download.
Find every occurrence before you miss one
The most common redaction failure after the overlay mistake is the missed instance — page 14 got the box, page 31 didn't. Type the name or number into Find & mark and the tool searches the PDF's entire text layer and boxes every occurrence at once. You review the marks, then flatten.
Private by construction
Redaction jobs involve exactly the documents you least want on someone's server. This tool runs entirely in your browser: the file is read into memory on your device, processed there, and the result is saved from there. The only network traffic is fetching the open-source pdf.js and pdf-lib libraries from a CDN. If you're on a plane with the file already open, it still works.
How to redact a PDF without Adobe Pro
The trap when you redact a PDF without Adobe Pro is the black box that isn't. Preview's Markup, the iPhone's markup pen and most free editors let you draw a filled rectangle over a name — and the name is still there. The rectangle is a shape sitting on top of the text. Select all and copy, search the file, or drag the box aside, and the text comes right back. That is how redaction failures in court filings keep happening.
Real redaction removes the content rather than covering it. Acrobat Pro does this with its Redact tool; this page does it by flattening. Draw your marks, or type the words and let Find & mark box every occurrence, then download: every page is re-rendered as a flat image with the marks burned in solid black, and a new PDF is built from those images. There is no text under the boxes because there is no text layer at all. Open the result and try to select under a box — there is nothing to select. The file never leaves your browser.
How to redact a PDF on Mac
You can redact a PDF on Mac with Preview — on the right macOS version. Since macOS 11 Big Sur, Preview has Tools → Redact: draw over the text and, when you save, Preview removes the content underneath rather than painting over it, and warns you that the change is permanent. That is real redaction. On older versions of macOS, Preview has no Redact item; the Markup toolbar's rectangle only draws on top, and the text stays in the file.
Two things to check even with the newer Preview. First, it does not search for you — if a name appears on fifteen pages, you find and cover all fifteen by hand. Second, Preview saves the redaction into the file you opened, so work on a copy. The tool on this page runs in Safari or Chrome on any Mac, finds every occurrence of a word for you, and produces a new flattened file while leaving the original untouched.
How to redact text in a PDF — and what redacting means
To redact text in a PDF means to remove it from the document for good, so that no reader, search, copy or file inspector can bring it back — not to hide it behind a shape. A redacted page should be safe to hand to the other side of a lawsuit, a journalist or a stranger.
On this page there are two ways to redact text. For a word you can name — a person, an account number, an address — type it into Find, step through every match with the prev and next buttons, untick any you want to keep, and mark them all at once. For anything else, drag a box over it; tick draw on every page for a header or footer that repeats. Zoom to 200% to place a box precisely.
Then verify, every time. Download the result, open it, and search for the word you removed: it should find nothing, because the pages are now images and the text layer is gone. Try to select under a box. If you can, you are looking at the wrong file.