View & remove PDF metadata

How to inspect and strip PDF metadata

1. Drop your PDF above. Its metadata fields appear the moment the file loads.
2. Edit, or strip. Change a field and hit Save, or use “Strip all metadata” to blank everything and delete the XMP stream.
3. Download the cleaned copy. Your original file is left alone.

The metadata you can't see in a viewer

A properties dialog shows you five or six fields. The file usually holds more. Most PDFs carry an XMP stream too, a block of XML the creating application embeds, and it can name authors, organizations, editing history and tool versions. No ordinary viewer puts that in front of you. This tool flags it, and “Strip all metadata” takes it out.

Where metadata actually hides

"The metadata" sounds like one thing. In a PDF it is at least two, stored in different places and written by different software, which is why files people believe they have cleaned often still identify them.

The document information dictionary is the old, simple one: title, author, subject, keywords, creator, producer, and the creation and modification dates. This is what a viewer's Properties dialog shows, and what most people mean by PDF metadata.

The XMP packet is the modern one: a block of XML embedded in the file, and usually far more talkative. Depending on what produced the document it can carry the author again, your organisation, the document's editing history, a unique document identifier that persists across saves and links revisions of the same file, and the exact version of the software used. Almost no viewer displays any of it. The two stores can also disagree, since clearing one in some applications leaves the other untouched, which is how a document with a blank Properties dialog still names its author to anyone who looks properly.

Stripping here clears the information dictionary and deletes the XMP stream outright, rather than blanking the fields a dialog happens to show. One niche caution: archival PDF/A files declare their conformance inside that XMP stream, so stripping metadata from one removes its claim to be PDF/A. If a court or an archive requires the format, clean the source document and re-export instead.

What metadata gives away

Any single field rarely matters much, but together they can be revealing. A CV exported from a work computer names your employer's licensed copy of Word, and a tender response can carry a creation date showing it was written three weeks before the invitation went out. An anonymous submission is authored by an account with your name on it. Redacted documents have been traced back to their source this way, repeatedly, and the document identifier in XMP is designed to connect versions of the same file: useful when you want it, and unhelpful when you have sent two supposedly unrelated documents to the same recipient.

Timestamps deserve their own mention: creation and modification dates are frequently the most sensitive fields in the file and the ones nobody thinks about.

What stripping metadata does not remove

Clearing metadata is one step; documents leak when it is treated as the whole job. Still in the file afterwards:

• Text hidden under drawn shapes. A black rectangle over a name is decoration; the name is still there and still copyable. That needs real redaction.
• Content cropped out of view. Cropping usually changes which part of the page is displayed, not what the page contains.
• Attachments, embedded files, and annotation authors. Comments carry the name of whoever made them.
• Image metadata. A photo placed in the document can keep its own EXIF data, including where it was taken.
• Earlier versions of the document, if it was saved incrementally; previous states can remain recoverable inside the file.

For a document going somewhere consequential, run the privacy scanner instead of guessing: it inspects all of the above and reports what it finds, so the decision about what to remove is an informed one.

Frequently asked questions