Redaction
The permanent removal of sensitive content from a PDF — including the underlying data, not just a visual overlay.
Redaction is the process of permanently removing sensitive content from a document so that no trace of the original information remains — not in the visible layer, not in the underlying data. The word comes from the Latin "redigere" (to bring back into order), but in modern usage it refers specifically to the censoring of text and images for disclosure purposes.
True redaction is a legal and compliance requirement in many industries. When a law firm produces discovery documents, a government agency releases records under a Freedom of Information request, or a hospital shares patient records with a researcher, sensitive information — names, addresses, financial figures, privileged communications — must be removed before disclosure. Getting it wrong has serious consequences.
The most common redaction mistake is using a black rectangle drawn on top of the text. This is a visual overlay — it covers the words on screen and in print, but the original text remains in the PDF's content stream and can be extracted by selecting the text underneath the shape, copying to a clipboard, or using a text-extraction tool. Genuine redaction requires that the underlying data itself be permanently deleted.
How proper redaction works
A correct redaction workflow involves three stages:
**Mark** — The redactor identifies the content to be removed. In a specialist tool, this means applying redaction annotations (a specific PDF annotation type with subtype /Redact) over the target areas. These annotations are markers, not yet permanent changes.
**Review** — Before applying, the marked areas are reviewed. At this stage the underlying content is still present and the annotations can be adjusted or removed.
**Apply** — The redaction is permanently applied. The software removes the content stream operators and data corresponding to the marked regions and replaces them with a filled black (or custom-coloured) rectangle drawn directly into the page content. After application, no software can recover the removed text or images — they no longer exist in the file.
The PDF specification defines a dedicated /Redact annotation type for exactly this workflow. Properly implemented, it ensures the original data is gone from the file, not just hidden.
Common redaction mistakes
Several widely-reported document disclosure incidents have resulted from improper redaction:
**Black box overlay** — Drawing a black rectangle or whiteout shape over sensitive text. The text remains in the content stream and is trivially extractable. This is not redaction.
**Highlighting in black** — Using a highlight annotation with a black colour. Annotations are a separate layer from content; the underlying text is untouched.
**Low-resolution printing** — Printing to PDF with the intent that the resolution will obscure text. Modern OCR and upscaling software can often recover content from poor-quality images.
**Metadata not cleared** — Even after correctly redacting page content, sensitive information may persist in the document's metadata, bookmarks, comments, attached files, or form field data. Complete redaction requires checking all document components.
Redaction versus whiteout
Whiteout and redaction are often confused because both result in a white or black area over content — but they are fundamentally different operations.
A whiteout places a coloured rectangle as an overlay on the page. The original content stream is untouched. Open the PDF in a text extractor, select text under the white box, or remove the overlay object, and the original content is fully visible. Whiteout is appropriate for cosmetic corrections — replacing a line of text you want to update — but is completely unsuitable for concealing sensitive information.
Redaction permanently modifies the page content stream. The underlying text and image data is deleted and replaced with a static rectangle. There is nothing left to extract. After proper redaction, the information is gone from the file — not hidden, not overlaid, gone.
Frequently asked questions
What is redaction in a PDF?
Redaction is the permanent removal of sensitive content from a PDF. Unlike covering text with a black box (which hides it visually but leaves the data in the file), true redaction deletes the underlying text and image data from the document's content stream so it cannot be recovered.
Is covering text with a black box proper redaction?
No. Drawing a black rectangle or using a black highlight over text is a visual overlay — the underlying data remains in the PDF and can be extracted by copying the text, removing the overlay object, or using a text-extraction tool. Proper redaction requires deleting the data from the content stream itself.
What is the difference between redaction and whiteout?
A whiteout is a coloured rectangle placed over content — the original data is still in the file, just covered. Redaction permanently removes the underlying data and replaces it with a static rectangle. Whiteout is for cosmetic edits; redaction is for permanent removal of sensitive information.
Do I need to clear metadata after redacting a PDF?
Yes. Page content redaction removes text and images from the visible pages, but sensitive information can still persist in metadata fields (author, title), bookmarks, comments, annotations, form fields, embedded files, and document history. Complete redaction must address all these locations.
Is redaction reversible?
No — that is the point. Once a proper redaction is applied and saved, the removed content cannot be recovered. This is why redaction tools have a two-stage workflow: mark for review first, then apply permanently. Always keep an unredacted original in a secure location if there is any chance you will need it later.