Remove the author from a PDF

The half most tools miss

• The Info dictionary — Author, Title, Subject, Keywords, Creator, Producer and the creation and modification timestamps. This is what "Document Properties" displays.
• The XMP metadata stream — a separate XML packet that duplicates most of the above. Clearing the Info dictionary alone leaves the author name readable in here.

We clear both. Note that removing PDF properties requires re-serializing the document, so unlike photos we cannot offer a byte-identical hash — and we say so on the result rather than implying otherwise.

What we deliberately keep

• Every page's content stream — text, images and vector graphics render exactly as before.
• Embedded fonts and colour profiles, so the document still prints and displays correctly.
• The document outline and tagged structure, so bookmarks and screen-reader accessibility survive the clean.

What a real dump looks like

A PDF's Info dictionary is small and easy to read even in a plain text editor — open a PDF in one and search for /Author or /Creator, and you'll usually find it in plain ASCII a few kilobytes into the file, uncompressed. The XMP packet is the opposite: a full RDF/XML document, often several kilobytes on its own, that duplicates Author, Title and CreateDate under completely different key names (dc:creator, dc:title,xmp:CreateDate) defined by Adobe's XMP specification rather than the core PDF spec. Clear one and not the other, and the document still identifies its author to anyone who knows to check both places.

The Info dictionary, the XMP mechanism, and the tagged-structure rules this page preserves are all defined in ISO 32000-2, the current PDF 2.0 standard. It formalised a lot of what Adobe's earlier PDF specifications did informally, including exactly how metadata streams and document structure are supposed to interact.

A detail most tools get wrong: what replaces a cleared date

Clearing CreationDate and ModDate sounds simple until you ask what value replaces them. Writing the current time seems reasonable and is exactly the wrong choice — it would swap one timestamp leak (when the document was created) for another (precisely when it was run through a cleaning tool), which is information you were trying to remove, not relocate. This tool writes a fixed epoch date instead, the same one on every file regardless of when it was actually cleaned, so the timestamp field is genuinely empty of information rather than quietly repurposed.

What survives common editors

"Print to PDF" from a word processor or browser generates a fresh Info dictionary from scratch — usually just Producer and a creation date, with no author unless the source document had one embedded in its own properties. Editing an existing PDF (adding a signature, merging pages, filling a form) with most desktop tools preserves the original Info dictionary and XMP untouched, because those tools are built to modify content, not metadata. Scanned documents are the quiet exception: many scan-to-PDF pipelines stamp the scanner's serial number or a default operator name into Creator or Producer, and that value stays identical across every document that scanner produces until someone reconfigures it.

Cleaning something else?

Scanned from a photo? Check JPG or PNG too if you still have the source image. For an Office document instead, see Word, Excel or PowerPoint.