Hidden Data in PDFs: Metadata and the Redaction Trap

When you share a document you are usually sending more than the pages you can see. A PDF can carry its author, its editing history, deleted text, and sometimes an entirely different document inside it.

What lives inside a PDF

• Document metadata: Author name, title, creation and modification dates, producing software.
• Your user name: Most programs fill the author field automatically from your OS account.
• File path traces: Some producers embed the full path of the source file.
• Comments and annotations: Present in the file even when hidden in the viewer.
• Layers: Hidden layers can be switched back on.
• Orphaned objects: Content deleted from the page but not from the file.

None of this appears on the page, all of it sits in the file, and all of it is accessible as text. Learning who prepared a document, in which folder and when, is often one "document properties" window away.

The redaction trap: a black rectangle hides nothing

Drawing a black rectangle over information is the most dangerous habit in PDF work. The rectangle is a separate graphic object drawn on top of the text. The text underneath remains in the file exactly as it was.

The consequence: anyone who selects and copies the text reads the redacted information directly. So do screen readers. So do text extraction tools. This has happened repeatedly with public documents and has caused real disclosures.

How to actually remove it

1. Delete the text, do not cover it. Remove the information from the document, then add a black box if you want the visual marker.
2. Rasterise the page. Once a page is pixels there is no text underneath. You lose searchability — which in this case is the point.
3. Remove the page entirely. If the sensitive material sits on one page, extracting the rest is the cleanest route.
4. Rewrite the document. An operation that builds the output from scratch, such as compressing or extracting pages, leaves orphaned objects behind.

A checklist before sharing

The third row is the most useful and takes thirty seconds: select all the text in the document and paste it into a plain text editor. Everything you see there is everything the recipient can see — redacted or not.

When this matters

These checks are not needed for every document. They should not be skipped for legal files, freedom-of-information responses, HR paperwork, medical records, reports containing customer data, and anything intended for public release.

If a document is leaving your organisation and contains information about other people, whoever shares it is responsible for the invisible content as well.

Frequently asked questions

More guides