What metadata is hidden in your PDF files
A PDF looks like a finished, sealed document. It feels like paper — fixed, final, nothing hidden. But every PDF carries a block of structured data that most people never see: the document information dictionary. It records who created the file, what software they used, and when they worked on it. Sometimes that is harmless bookkeeping. Sometimes it is a problem.
What the metadata fields contain
The PDF specification defines a standard set of metadata fields in the document information dictionary:
- Author — the name configured in the authoring software. In Microsoft Office, this defaults to the account name. Many people never change it, so documents attributed to "Anonymous" or a pseudonym quietly carry their real name.
- Creator — the application that produced the original content. "Microsoft Word 2021" or "Google Docs" or "LaTeX via dvipdfmx."
- Producer — the software that generated the PDF itself. Often a library name and version number.
- CreationDate and ModDate — timestamps recording when the document was first created and last modified, down to the second, including timezone offset.
- Title and Subject — usually whatever the author typed, or the filename by default.
- Keywords — occasionally populated, more often empty.
Beyond the dictionary, many PDFs also carry an XMP packet — Adobe's XML-based metadata format. XMP can store everything the dictionary holds, plus revision identifiers, editing history markers, and the unique document ID that Adobe applications generate for every file.
When it has caused real problems
The author field is the most common source of trouble. A whistleblower submitting a document anonymously, a lawyer sending a draft that should not be attributed to a specific person on the team, a journalist receiving a leaked file — in each case, the Author field can undo the anonymity that the sender assumed they had.
Timestamps tell their own story. A document claimed to have been created on a certain date, but the CreationDate field shows it was actually produced two weeks later. In legal proceedings, metadata timestamps have been used to challenge the authenticity of documents and to establish when edits were actually made.
The Creator and Producer fields reveal the toolchain. A government report distributed as an official PDF, examined more closely, turns out to have been created in a pirated copy of an old Adobe version — an embarrassment, not a catastrophe, but the kind of detail that erodes credibility.
How to view it
In Adobe Acrobat, File → Properties shows the document information dictionary in a clean panel. Most PDF readers have an equivalent.
On the command line, exiftool document.pdf dumps every metadata field including XMP. The pdfinfo utility (part of Poppler) shows the dictionary fields in a single line each.
In a browser, this site's PDF tool parses the file locally and displays the fields before offering to remove them.
How to remove it
In Adobe Acrobat Pro, the "Remove Hidden Information" feature under the Protection menu strips metadata, comments, hidden layers, and embedded objects. It is thorough but requires a paid licence.
On the command line, exiftool -all= document.pdf clears the dictionary and the XMP packet. For a free graphical tool, MetaScrub's PDF metadata remover does the same in the browser — no upload, no software to install.
One important note: removing metadata from a digitally signed PDF invalidates the signature. That is by design — the signature covers the entire file, and any modification breaks it. If the signature matters, you need to re-sign after cleaning.
A habit, not a tool
The right time to think about PDF metadata is before you send the file, not after. It takes ten seconds to check the Author field. If you are distributing documents where attribution matters — or where it should not exist — make it part of your workflow. The tool is simple. The habit is the hard part.