Every PDF you send carries more than its pages. Next to the text and images sits a set of document properties, and often an XML metadata packet, that can name the person who wrote the file, the program they used, and the exact minute and time zone it was saved. Before you share a CV, a contract or anything sensitive, it is worth taking two minutes to check those fields and remove metadata from a PDF you would rather keep anonymous. This guide explains what the hidden fields reveal, how to read them, how to clean or edit them in your browser, and what cleaning does not change.
What hidden metadata a PDF carries
Most metadata lives in two places. The Info dictionary holds the classic document properties you see in a reader's Properties dialog. The XMP packet is a block of XML that repeats many of the same values and can add more, such as document IDs and editing history. A file can contain both, and they do not always agree.
| Field | Where it usually comes from | What it can reveal |
|---|---|---|
| Author | The user name set in the authoring app or operating system | Your full name, a login name, or the name of whoever set up the laptop |
| Creator | The app the original document was made in | That a “scanned original” was really typed in a word processor |
| Producer | The library or driver that wrote the PDF | Which converter, print-to-PDF driver or editing tool touched the file |
| CreationDate / ModDate | The clock of the computer that saved the file | When the file was made and last changed, including the time zone |
| Title, Subject, Keywords | Often left over from a template | An old client's name or an internal project name |
| Custom properties | Document management systems and office suites | Case numbers, department codes, internal IDs |
| Attachments | Files embedded in the PDF | File names and sometimes local folder paths with a user name in them |
Dates come with a time zone
PDF dates are stored as a compact string that includes the offset from UTC. A creation date like this one tells a reader not only when the file was made, but roughly where:
D:20240115103000+05'30'
→ 2024-01-15 10:30:00 +05:30That +05'30' offset narrows the author's location to a small part of the world. For most documents this is harmless, but for an anonymous report it can matter more than the author field.
Old copies left by incremental saves
The PDF format lets an editor save changes by appending them to the end of the file instead of rewriting it. The previous version of each changed object stays in the file, just no longer used. If you changed the Author field in an editor that saves incrementally, the old author name can still be sitting earlier in the same file, readable by anyone who opens it in a text editor.
Why it matters before you share a PDF
- CVs and cover letters. A resume built from a template may still carry the template author's name, or a title such as the name of the last company you applied to. Anyone who opens the document properties sees it.
- Contracts and proposals. A Title or Subject left from a previous client's version reveals that the document was reused. A modification date that is later than the date written on the page invites questions.
- Anonymous submissions. Whistleblower reports, peer reviews and public comments are often meant to be anonymous. The author name, the authoring app and the time zone together can be enough to identify someone.
- Internal files going public. Custom properties added by a document management system can expose internal case numbers or server paths.
How to check a PDF's metadata first
Start by looking at what is actually in the file. Drop it into the PDF metadata viewer, which reads everything on your device without uploading it. It shows:
- Title, Author, Subject, Keywords, Creator and Producer, with identifying fields such as the author, creator, company and attachment file paths highlighted.
- Creation and modification dates with the time zone written in the file, next to the raw date string.
- Custom properties and the full XMP metadata.
- Whether the file is encrypted and which actions its permissions allow or block: printing, copying, editing.
- Page sizes in points, millimeters and inches, with paper names like A4 or Letter.
- Whether the document is tagged and contains forms, signatures, bookmarks, attachments or document JavaScript.
If the PDF is password-protected, type the password to unlock it. The password is used only in your browser. Compare the Info fields with the XMP section: a file whose Author was edited in one place but not the other still gives the old value away.
How to remove metadata from a PDF in your browser
Open the same file in the PDF metadata remover. It lists the properties it found, then gives you two options.
- Drop the PDF onto the page and read the summary of properties and XMP metadata.
- Click Remove all metadata. This removes the whole Info dictionary, including custom keys, every XMP packet in the file, the
/PieceInfodata that editing apps store for themselves, and objects that are no longer referenced, which covers the old copies left by incremental saves. - Read the result. The tool re-reads the saved copy and shows what is left, so you are not relying on a promise.
- Download the cleaned PDF and share that file, not the original.
Editing the fields instead
Sometimes you want a clean file that still has a sensible title. Open Edit the fields instead to change Title, Author, Subject, Keywords, Creator and Producer. Leave a field empty to delete it. Two checkboxes, on by default, also remove the created and modified dates and any custom properties. The XMP packet is always removed when you edit, because it would still contain the old values.
A practical setup for a proposal: set the Title to Website redesign proposal, set Author to your company name instead of your own, clear Creator and Producer, and keep the date boxes checked. Then click Save edited copy and check the summary.
What removing metadata does not change
Metadata cleaning is narrow on purpose. Know its limits before you rely on it.
- Visible content stays. Pages, text, images, links and form fields are unchanged. A name in the footer, a signature block or a letterhead is content, not metadata.
- Covered-up text stays. A black rectangle drawn over a sentence hides it on screen, but the text underneath is still in the file. Run the cleaned PDF through PDF to Text: if the “redacted” words appear in the output, they are still there. Use a real redaction feature in your PDF editor instead.
- Digital signatures become invalid. Any change to a signed PDF breaks the signature. Clean the file before it is signed, never after.
- Encrypted PDFs are refused. Saving an encrypted file this way would drop its password protection, so the remover does not change it. Remove the password in the app that set it, clean the file, then protect it again.
- Attachments stay. If the viewer lists an embedded file you do not want to share, delete it in your PDF editor or export the document again without it.
Export cleaner PDFs from the start
Most PDF metadata is inherited from the source document, so cleaning at the source saves a step every time.
- Clean the Word, Excel or PowerPoint file first. The PDF export often copies the title and author from the source. The Office metadata remover strips author, last modified by, company, template path and custom properties from .docx, .xlsx and .pptx files before you export. In Office for Windows, File → Info → Check for Issues → Inspect Document does a similar job.
- Turn off document properties in the export. In Word for Windows, the Options button in the Save As PDF dialog has a Document properties checkbox under non-printing information. Unchecking it leaves the document properties out of the exported PDF.
- Set your app's user name deliberately. The Author field usually comes from the name in your app's settings. Change it once and new documents start out right.
- Prefer a full save for final versions. When an editor offers a choice, saving to a new file rewrites the PDF and leaves fewer old revisions behind.
A quick checklist before you hit send
- Open the final PDF in the metadata viewer and note anything identifying.
- Remove or edit the fields, and confirm the summary of what is left.
- Extract the text to make sure nothing hidden under shapes or white text survives.
- Sign the document only after it is clean.
- Share the cleaned copy, and delete any older versions from shared folders.
The whole routine takes a couple of minutes and runs on your own device, which is the right place to handle a file you are trying to keep private.