Skip to content
CreatorPrivacyKit
Learn Blog Privacy @mollycoddle330

PDF Info dictionary and XMP remover

PDF Info dictionary and XMP remover?

This page reads the PDF document information dictionary and the catalog Metadata stream in your browser, writes a new file without those values, and reads that download again so you can see what remains, while page text, forms, annotations, attachments, and signatures stay in the file.

Last updated

This page reads the PDF document information dictionary and the catalog Metadata stream in your browser, writes a new file without those values, and reads that download again so you can see what remains, while page text, forms, annotations, attachments, and signatures stay in the file.

Last checked 2026-10-22

Read, clear, read again

No file chosen. Encrypted files are not modified.

Before

After

What one generated PDF showed

A one-page PDF written with pdf-lib in this session was 1331 bytes and carried author GapPdfAuthor plus an uncompressed catalog Metadata stream containing the marker GapXmpAuthorCPK-unique-marker. After the clear, the saved file was 460 bytes, the marker and the author string were absent, a fresh parse reported no XMP and no author, title, or producer, and the page count was still 1. The verify diff of those two parses was clean-of-readable-metadata.

That file was created here. ExifTool’s PDF tag page says its own edits are incremental updates and that old information is never actually deleted until a later step such as qpdf --linearize. This page does not claim every other producer drops the old bytes the same way. It claims what the re-read of this rewrite showed.

The Info-only remover still leaves XMP. The PDF metadata viewer reads and does not edit. The author guide separates the Info author from a name printed on the page.

Sources

Last checked 2026-10-22.

Questions

Does this upload the PDF?

No. pdf-lib rewrites the file in this tab and the page reads the new bytes again.

What is deleted?

The standard Info keys Title, Author, Subject, Keywords, Creator, Producer, CreationDate, ModDate, and Trapped, plus the catalog Metadata entry and the stream it pointed at when that stream was an indirect object.

What stays?

Page text, forms, annotations, attachments, JavaScript, and signatures are not part of this clear. An encrypted PDF is refused and is not unlocked.

Is this the same as the other PDF remover?

No. The PDF metadata remover clears Info fields and leaves the catalog XMP packet. This page deletes that packet as well on files pdf-lib can load.

Related pages