How to redact a PDF permanently

Redacting is not the same as covering

A black rectangle drawn over text in a PDF editor hides it from view and nothing more. The text is still in the file, and anyone can select it, copy it or find it with a search — which is how redacted documents leak.

Redact PDF removes it instead. The box is not laid over the text. With “Flatten every page”, each page is rendered to a JPEG image, the box is painted into that image's pixels, and a new PDF is built from the images alone. None of the original's text, drawings, form fields or annotations is copied into it, and neither is its document information — the new file's only title is “Redacted PDF”.

Flatten every page, or remove only the text

With “Flatten every page”, the method selected when the tool opens, every page — not only the ones with a box — is rebuilt as a flat JPEG image, so the result has no selectable or searchable text, links, form fields, annotations or bookmarks on any page.

With “Remove text under boxes”, only what lies under each box is taken out of the file — its text, the covered pixels of any picture and any annotation there — and the rest of every page stays as it was, still selectable and searchable, along with the file's bookmarks and properties. Vertical text, a form field or certain images under a box cannot be removed this way, and the redaction then stops with a message instead of saving.

If you flattened and the rest of the document needs to be searchable again, run the redacted file through OCR PDF afterwards. OCR recognizes the text that is still visible on the pages; the text under a box is gone, so there is nothing there for it to find.

When flattening, choose the resolution before you run: Screen (96 DPI) or Print (150 DPI). A higher resolution keeps small print legible and makes a larger file.

Find every instance before you redact

Search the PDF's text, with Match case and Whole word, or detect email addresses, phone numbers, payment card numbers, IBANs and dates of birth — card numbers only when they pass the card checksum, IBANs only when they pass the IBAN checksum, and dates of birth only after a “Date of birth” or “DOB” label. Every match is listed with the words around it to tick or untick before anything is redacted. A page with no selectable text, such as a scan, has nothing to find.

Detection matches patterns and checksums, not meaning: it does not know a name from any other word, and it can miss a number written in a format it does not expect. Treat the list as a starting point, search for the names yourself, and read the list before you redact.

Your file stays on your device

Redact PDF is an in-browser tool: your browser renders the pages and rebuilds the file, so the document you are redacting never leaves your device. That matters most for exactly the files people redact — contracts, statements and records with personal details in them.

It is not fully offline. The page and its processing engine load over the network, and parts of the engine are fetched the first time a step needs them. The processing itself runs on your device. Rendering a page that holds a JPEG 2000 or JBIG2 image, as scanned documents often do, downloads an image decoder from our site the first time it is needed, and that step needs a connection. Chinese, Japanese or Korean text whose font is not embedded in the PDF needs a character map, downloaded from our site the first time it is needed, and that step needs a connection.

Before you share it

Open the downloaded PDF and check that each box sits where you meant it to before you send the file anywhere: an area typed in by position that sits a little off leaves the edge of a word readable.

Redaction removes content; it does not lock the file. If the redacted PDF should also need a password to open, run it through Protect PDF afterwards.

• How to password-protect a PDFProtect PDF
• How to repair a damaged or corrupted PDFRepair PDF
• How to compare two PDF filesCompare PDF