PDF redaction guide
Hidden text, metadata and attachments in redacted PDFs
PDF is a container, not a flat screenshot. Page content, annotations, form fields, metadata and embedded files can coexist independently, so a visible black cover addresses only one layer.
Extractable text
Text may remain selectable or searchable beneath a visual annotation. Region-based extraction can confirm some failures, while image-only pages and custom encodings remain an explicit unknown.
Document metadata and XMP
Title, Author, Subject, Creator and Keywords can reveal names or project context. Producer and timestamp values are also worth inventorying, even though macOS may regenerate technical values when writing a copy.
Comments and form values
Text notes, free-text review comments, stamps and filled Widget values may contain internal discussion or personal data outside the visible page narrative.
File attachments
A FileAttachment annotation or EmbeddedFiles name tree can carry a complete source document inside the PDF. Both locations need inspection before distribution.
Unsupported means unknown
No single parser observes every PDF construct. A defensible workflow records which structures were inspected and escalates unsupported objects instead of issuing an all-clear.