Privacy & safety
How to Remove Metadata From a PDF (And What Survives It)
A PDF usually remembers who made it, with what, and when. Here is how to remove that, and the one thing that stays visible even inside an encrypted file.
Zahid Sheeraz5 min readExport a PDF and it takes a small set of notes about itself. Who made it, what it is called, which company it belongs to, which program produced it, and the exact moment it was last changed. None of that is visible when you open the file - it is tucked into a part of the document called the document information dictionary, and it shows up in the properties panel of your file manager or your PDF reader.
It is also, in a lot of cases, the only thing that identifies you in a file you thought was anonymous.
What a PDF remembers about you
Open the properties of any PDF you have made and you will find some subset of this:
| Field | What it usually contains |
|---|---|
| Title | The document name, often the first line of the title |
| Author | The person who created it - the account name in Word, your full name in Acrobat |
| Subject | Sometimes a summary, sometimes a file path |
| Keywords | Words that were in the file's properties when it was saved |
| Creator | The program that made the original - "Microsoft Word 16", "Adobe InDesign 2024" |
| Producer | The library that wrote the file, sometimes with a version |
| Creation / ModDate | When it was made, and when it was last saved |
Documents saved by modern tools often carry a second block: XMP metadata, a stream that holds richer information - including the name of the last person who saved the file, the editing time, and the document history. Acrobat calls it "Document Properties" and shows the same data in two places for a reason.
None of this is a security feature. It is simply what the format records while it saves.
How it leaks in practice
The "anonymous" export that names you. You strip the author field before sending, then export to PDF, and the export writes your account name back into the new file. The output is cleaner than the input, but not clean.
A document prepared for someone else. A quote, a pitch, a proposal: the file says "Prepared by Jane Doe, Northgate Ltd" in its properties, and the recipient can see that without opening it. If the relationship has not started yet, that is information you handed over for free.
A file that used to belong to someone else. You edit a template, a contract or a report somebody sent you. The author field still says them. Now the file identifies a third party to your recipient, who may not know they are being named.
A shared drive, a public folder, an upload. Metadata is a quick way to work out who a leaked document came from - which is exactly why it is worth removing before you publish rather than after.
Your own toolchain fingerprints you. The Creator field tells a reader which software made the file, which version, and sometimes on which platform. That is not identifying on its own, but combined with the Author field it is.
What removing metadata does - and does not - touch
The Remove PDF Metadata tool deletes the document information dictionary and the XMP stream. It reports exactly which fields it found, so you know what was there.
What it does not do:
- It does not change the content. A document that says "Jane Doe" in the second paragraph still says it. Metadata is the file's label, not its text.
- It does not remove annotations. Comments and notes carry their own author names, stored with the annotation rather than in the document properties. A redacted-looking comment signed "Jane Doe" is still there.
- It does not clear form fields. The values typed into a form are part of the page content, not metadata.
- It does not remove embedded files. A PDF can carry another file inside it, with a name you can read.
- It does not remove revision history or tracked changes in a document that still has them.
So it is one step of a cleaning process, not the whole process. For a document you are about to send to a stranger, the realistic order is: strip the metadata, and check the content and the annotations yourself.
The detail that catches everyone: encryption does not hide metadata
This is the part worth knowing, because it reverses the order you would naturally choose.
The document information dictionary in a PDF sits in the trailer, outside the encrypted content. When you password-protect a file, the pages and the strings inside them are encrypted - but the properties panel still shows the author, the title and the producer. Anyone who receives the encrypted file can read its metadata without the password. They just cannot read the document.
There is a second consequence, and it is the practical one: you cannot strip metadata from an encrypted file without the password. The metadata is outside the encryption, but the tool that would remove it still has to open the file to do the work.
So the order is not a matter of taste:
- Remove the metadata first, while the file is still open.
- Then encrypt it, if it needs a password.
- Then send it, with the password on a different channel.
Doing it the other way round means either stripping before encrypting anyway, or unlocking the file first to clean it - which is more work and more risk than it sounds.
This is the pair to use together: Remove PDF Metadata then Protect PDF. One removes what identifies the file, the other stops the file being read. Neither does the other's job - and the order between them matters.
A short checklist
- Look at the properties first, so you know what you are dealing with.
- Strip the metadata while the file is still unencrypted.
- Encrypt afterwards - and remember the metadata would have been visible either way.
- Check the content and the annotations yourself. Metadata is the label, not the text.
- Expect the tool to tell you what it removed - an empty list means there was nothing there, which is worth knowing too.
- Keep the original. The cleaned copy is a new file.
- Re-check after every export. Saving to PDF again can write the metadata straight back.