How to redact a PDF properly
Redaction is a short job done carefully rather than a long job. The steps below take a few minutes on a normal document, and the last one — verifying the output — is the step that separates a redacted file from a file that looks redacted.
Work from the original, and keep it
Start from the source PDF rather than a printout, a scan of a printout or a screenshot. Each of those conversions can introduce artefacts and none of them help. Keep the unredacted original somewhere separate and clearly named: you will need it if a question comes back, and the redacted copy is not a substitute for a record. Never overwrite the original with the redacted version — that is a mistake nobody notices until the day they need the original.
Mark every occurrence, not the obvious one
Open the document and mark what has to go. The name in the heading is easy; the same name in the signature block, in a footnote, in an email address, in a file path in a footer, and in the document's own metadata is where redaction fails. Use the search to find every occurrence of each term rather than reading page by page — human attention is the weakest part of this process. If a page contains a table or a scanned image, remember that a name inside an image is not text, so searching will not find it: you have to see it.
Redact, then check what you produced
When the boxes are placed, produce the output. The app rasterises each page that carries a box, paints the boxes onto the raster, rebuilds the page from it, and copies untouched pages across as they were. Then — and this is the part most tools skip — it reopens the file it just built and searches it for the terms you removed, twice over: once by extracting text the way a copy-paste would, once by scanning the raw streams the way someone examining the file would. If either search finds a term, you are told before you download, not after you send.
Before you send it
Three habits worth keeping. Open the output yourself, select the redacted region and try to copy it — it takes five seconds and it is the check the recipient will do. Look at the document's properties: title, author and subject fields survive from whatever produced the file and often carry a name, a client, a matter number or a full file path. And send the redacted file as a new file with a new name; sending a file called `contract-final.pdf` that is now a different document is how the wrong version gets forwarded.
Frequently asked questions
Do I need to install anything?
No. The redaction, the rebuild and the verification all run in the browser tab. The file is not uploaded.
Can I redact a scanned document?
You can cover regions of a scan, and since a scan is already an image there is no text underneath to remove. But be aware that if the scan has been through OCR, there may be a text layer, and that layer has to go too — which the verification step will tell you.
What happens to pages I did not redact?
They are copied from the original unchanged, so they keep their vector text and stay selectable and searchable.
Can I redact part of a word or a single digit?
Yes, you mark a region rather than a word. Bear in mind that a partly visible identifier is often enough to reconstruct the whole thing.
Does the app keep a copy of my document?
No. Everything happens in the tab; there is no upload and nothing to store.
What if the verification finds something?
It tells you which term it found and where, so you can go back and mark the occurrence you missed. That is the entire point of running it.
Remove the text, then prove it is gone. Redact takes the words out of the page rather than covering them, then reopens the file it just made and searches it twice — by text extraction and by raw stream scan — before letting you download. The file never leaves your browser.
Redact a PDF in the browser