Home / Guides / PDF metadata explained
PDF metadata explained — the resume leak, the ghost revisions, and the fix
Photos carry EXIF; documents carry something worse: your name, your company's software licenses and sometimes the drafts you thought you deleted.
Last updated: · ~7 min read
The two hiding places
A PDF keeps its administrative secrets in two structures. The document information dictionary is a small table referenced from the file trailer: Title, Author, Subject, Keywords, the creation and modification timestamps, and two fields — Producer and Creator — that name the software chain that produced the file. The XMP metadata block is a second layer: an entire XML document embedded inside the PDF, where modern apps store richer records such as the contributor list, the original document title and tool-specific history.
Neither is exotic. Office suites fill them by default: export a CV from Word and the Author field is populated from the account name of whoever was logged in — full name, sometimes employer. WPS and Pages behave the same. The recipient needs nothing more than a Properties dialog to read it.
Where it bites: three familiar scenarios
The résumé. Job seekers routinely anonymize the visible text of a CV for agencies — then export to PDF and mail a file whose properties still name them. Recruiters' parsing software reads that field automatically. The manuscript. Double-blind peer review has been broken more than once by author details surviving in the PDF. The tender. Bidding documents expose editing timelines and software lineages that competitors and auditors can, and sometimes do, inspect.
The ghost revisions problem
PDFs have a trait photos don't: when a file is edited and saved again, many tools write an incremental update — new objects appended after the old ones rather than a full rewrite. The old version stays physically inside the file. Delete a page, black out a paragraph, rename the author — and in many cases the original content is still buried in the file, recoverable with nothing more exotic than a text editor. Cases have surfaced of "redacted" court filings and reports whose secrets were readable underneath.
A proper cleaning tool therefore has to do more than blank a few fields: it must rebuild the file from its live structure only, discarding every object the current version doesn't reference. That's exactly what our PDF cleaner does — metadata fields, XMP block and orphaned revision remnants all removed, visible content untouched, processing entirely in your browser.
How to check and clean your own documents
- Inspect first: drop the file into the PDF metadata viewer — the full field list appears locally, nothing is uploaded.
- Anything you wouldn't hand to the recipient? Strip it — one pass, batch supported, download as clean copies.
- For new exports, clear the fields at the source: Word → File → Info → Inspect Document → Document Properties and Personal Information → Remove All, before exporting to PDF.
- Remember what's visible is not metadata: comments, tracked changes and filled form fields stay in the document by design — clean those in the editor itself.
Photos have the same story with different field names — start at what EXIF data is if you also share images.
Frequently asked questions
Does "Print to PDF" remove metadata?
Mostly yes — it produces a brand-new file through the print pipeline, so the original author fields don't survive. The trade-offs: text can be re-flowed or rasterized (searchable text and selectable quality may suffer), bookmarks and links are lost, and the new file still carries its own Producer field naming your PDF printer. A dedicated metadata cleaner keeps the original quality and simply removes the fields.
Where do I see a PDF's metadata myself?
Adobe Acrobat: File → Properties → Description. Free readers show parts of it, and our browser-based viewer lists the full field set — Info dictionary plus XMP — without uploading anything: the tool lives here (photo and PDF inspection share the same privacy model).
Can metadata be used to track who leaked a document?
Yes — that's exactly why organizations care. Author fields, incremental edit histories and hidden revisions have been used to trace leaked files back to individuals and to identify the software and machine lineage that produced them. The same trail that protects organizations can expose individuals who never knew it existed.