The parts of a document nobody reads but everybody can
A PDF carries a quantity of information that never appears on any page. Your name, the software you used, the file path it was saved from, when you started and finished it, and sometimes fragments of earlier drafts. None of it is hidden in any meaningful sense — a text editor or any PDF tool will show it. It has embarrassed law firms, exposed anonymous sources, and revealed which template a supposedly bespoke proposal was built from. This guide covers what is actually in there and how to remove it before a document leaves your hands.
What is stored
| Field | Typical value | Why it matters |
|---|---|---|
| Author | Your full name, or a colleague's | Reveals who really wrote a document signed by someone else |
| Title | Often the original filename, not the document title | 'Proposal - reuse from Acme - v3' is a real example of the genre |
| Producer / Creator | Word 16.0, LaTeX, Canva, a specific accounting package | Reveals tooling, versions and sometimes the organisation |
| CreationDate / ModDate | Exact timestamps with time zone | Shows a document dated Monday was written Sunday night |
| Keywords / Subject | Whatever a template had | Frequently inherited from a file someone copied |
| XMP metadata | A second, richer copy of all of the above | Removing only the basic fields leaves this behind |
| Embedded file paths | C:\\Users\\firstname.lastname\\Clients\\... | Leaks a real name and the folder structure of a business |
Two things make this worse than it sounds. The metadata is duplicated: PDFs store an Info dictionary and, usually, an XMP packet with overlapping fields, so a tool that clears one and not the other leaves the data in place. And it is inherited: merging documents, converting from Word, or starting from a template all carry metadata forward from files you may not remember.
What else travels invisibly
- Annotations and comments. Review notes are stored as objects and are frequently invisible in a quick preview but plainly visible in another reader.
- Form field values, including fields scrolled out of sight or set to hidden.
- Content outside the crop box. Cropping a PDF usually changes the visible window rather than deleting content. The material beyond the edge is still in the file, and uncropping brings it back.
- Layers turned off. An optional content group that is not displayed is still present.
- Image metadata. Photographs placed into a PDF can retain their own EXIF, including GPS coordinates.
- Earlier revisions. PDFs support incremental saving, which appends changes rather than rewriting the file. Previous states can remain recoverable.
Black rectangles are not redaction
This is the failure that produces news stories. Drawing a filled black box over a name in a PDF editor adds a rectangle to the page. The text underneath is untouched: select across the area and copy, or extract the text layer, and the name is right there. The same applies to white boxes, to highlighter set to black, and to covering text with an image.
Genuine redaction removes the underlying content, not just the view of it. If a document requires real redaction — legal disclosure, a freedom of information response, anything involving a third party's personal data — use software built for it and verify afterwards by extracting the text and searching for what should be gone. For lower-stakes cases, rendering the pages to images and rebuilding the PDF removes the text layer entirely, at the cost of searchability and file size.
A cleaning routine before sending
Four steps, in this order, cover nearly everything:
- 1. Flatten. Flatten PDF fixes form values and annotations into the page and removes the live objects that carried them.
- 2. Remove metadata. Remove PDF Metadata clears author, producer, timestamps and the XMP packet. Use Edit PDF Metadata instead when you want a deliberate title and author rather than an empty one — a blank author field on a formal document looks odd.
- 3. Rebuild the file. Compact PDF writes a fresh document, which discards unreferenced objects and any incremental-save history rather than appending to it.
- 4. Verify. Extract the text with PDF to TXT and search it for anything that should not be there. This takes thirty seconds and is the only step that actually confirms the rest worked.
When to keep metadata
Stripping everything is not always right. Metadata is what makes a document findable in a document management system, and archival standards expect it — PDF/A requires consistent metadata precisely so that files remain identifiable decades later. For published reports and public documents, set the title, author and subject deliberately rather than clearing them. The rule is not remove everything; it is know what is there and choose.