Remove PDF Metadata
Remove PDF Metadata shows and strips the properties a PDF carries about you: the author, title, subject, keywords, the software that wrote it and the timestamps. It clears both the Info dictionary and the XMP packet most tools leave behind, then reads the result back to check the values are gone.
Drop a PDF here
or drop anywhere on this page · Choose a file · 100 MB
Accepted: PDF
Drop a PDF to see what it carries about you.
How does Remove PDF Metadata work?
A PDF carries document properties in two places, and a tool that cleans one of them has not cleaned the file. Knowing which is which is most of the work.
The Info dictionary, which is the old way
A plain dictionary in the trailer, with /Author, /Title, /Subject, /Keywords, /Creator, /Producer, /CreationDate and /ModDate. Every reader shows these in a Properties dialog, and they survive almost every editor untouched — a document written in Word in 2019 and re-exported six times still carries the original author.
XMP, which is the new way and the one people forget
An XML packet in a stream the catalog points at. It repeats the same fields in a different format and adds whatever the producing application felt like recording: an editing history with the original author, the camera model, sometimes GPS coordinates. Word, Acrobat and most print workflows write both, and the two can disagree. Removing the Info dictionary alone leaves the XML packet sitting in the file with the author ’s name in it — which is the failure this tool exists to avoid.
Deleting the reference is not enough either. The packet is an object, and the library that writes the output writes every object it knows about — so the reference goes and the author’s name stays, as an orphan nothing points at and every text editor can still read. It has to be unregistered, not just unlinked.
Why the result is read back before you are told it worked
The library used here stamps its own /Producer and a fresh /ModDate every time it writes a file, so those two fields cannot simply be deleted — they come back. The output is therefore patched at the byte level, and because every PDF string is delimited by its own parentheses, the contents can be overwritten with spaces of exactly the same length. Every cross-reference offset in the file stays where it was, which is what makes the edit safe.
Afterwards the tool reads the output again and looks for the values it was supposed to remove. If any survived, it says which. A tool that claims to have removed a person’s name and has not is worse than one that admits it could not.
What this cannot remove
Metadata is not the only thing in a document that identifies you. Text on the page, an email address in a footer, a signature image, a comment left in a margin, the filename itself — none of that is metadata and none of it is touched. Embedded fonts keep their own names, which sometimes include the foundry. And the document’s own structure can reveal the software that made it even with every field cleared. For a document going to a stranger, read it as well as cleaning it.
Before
/Author (Ada Lovelace) /Title (Q3 acquisition — Project Falcon) /Creator (Acme Writer 4.2) /Producer (Acme Writer 4.2.1) /CreationDate (D:20260911143012Z) + an XMP packet repeating all of it
After
/Author () /Title () /Creator () /Producer ( ) /CreationDate ( ) /ModDate ( ) no XMP packet
What options and edge cases does Remove PDF Metadata support?
| Parameter | Type | Default | Behaviour & edge cases |
|---|---|---|---|
| Author | Info | removed | The account name of whoever created the document. Usually a real person and almost always the field most worth removing. |
| Title | Info | removed | Often the original filename, which frequently contains a client name, a project code or a matter number. |
| Subject | Info | removed | Free text. Sometimes a description, sometimes an internal ticket reference. |
| Keywords | Info | removed | Comma-separated tags, often carrying project codes that identify what the document is about. |
| Creator | Info | removed | The application that authored the content, before the PDF was produced. Names your toolchain. |
| Producer | Info | blanked | The application that wrote the PDF, usually with a version number. Cannot be deleted — the writing library re-adds its own on save — so it is overwritten with spaces instead. |
| CreationDate / ModDate | Info | blanked | When the document was written and last saved. A timestamp plus a filename narrows down a timezone and a working day. |
| XMP | stream | removed | The XML packet: the same fields again, plus whatever the producing application added. The part most metadata removers miss, and the part that most often holds a GPS position. |
| Page content | not touched | unchanged | Text, images, links, form fields, annotations and bookmarks are all preserved exactly. Only the properties change. |
| Verification | checked | always | The output is read back and searched for the values that were supposed to be gone. Anything that survived is named rather than quietly ignored. |
Frequently asked questions
Why does the output still have a Producer field?
Because the library that writes the PDF stamps its own Producer and a fresh ModDate on every file it saves, and there is no way to ask it not to. The value is overwritten with spaces of exactly the same length instead, so the field exists and is empty. Readers show a blank property rather than the tool that wrote it — which is the outcome that matters.
Does removing the metadata change the document?
No. The pages, the text, the images, the links, the bookmarks and any form fields are preserved byte for byte; only the properties dictionary and the XMP packet change. What you lose is the original creation date and the author, which is the point. What you keep is everything a reader would notice.
Is this the same as redaction?
No, and confusing the two is dangerous. This removes the properties describing the document. Redaction removes content from the page — a name in a paragraph, an account number in a table. If the identifying information is visible when you read the page, metadata removal does not touch it and you want the redaction tool instead. Doing one and believing you did the other is how documents leak.
What about metadata in the images inside the PDF?
Not covered. A photograph placed on a page keeps its own EXIF data — camera, timestamp, sometimes GPS — inside the image stream, and this tool does not decode or rewrite those. If the images matter, run them through the EXIF remover first and rebuild the PDF from the cleaned copies.
Can the metadata be recovered afterwards?
Not from the output. The values are overwritten in place rather than unlinked, so there is no orphaned copy left anywhere in the file for a recovery tool to find. The original file on your disk is untouched, of course — this writes a new one and never modifies what you dropped in.
Is my PDF uploaded?
No. The document is parsed, rebuilt and verified inside a Web Worker in this tab — open the Network panel and drop a file to see that nothing leaves. That matters here more than anywhere else on this site: a document whose metadata you are removing is usually the one whose metadata is the sensitive part.