scriptiii
100% Client-Side

Info dict · XMP · no branding

PDF metadata remover

Read the document properties and XMP packet a PDF carries — including the tool it names — then save a copy with all of it deleted and nothing stamped in its place.

The tool reads and rewrites the file in this tab. Your document is never uploaded.

Short answer: drop the PDF to see its Title, Author, Creator, Producer, dates and any XMP packet. Press remove and the saved copy has no property fields, no XMP and no packet naming a tool — and the page re-reads that saved file to show it.

Where the AI tool’s name comes from

Two places, and this tool clears both. The /Info dictionary holds Creator and Producer strings that the export pipeline writes — often a printer or app name, sometimes the assistant itself. Separately, an XMP packet can hold a claim generator, an editing history and a creator tool. Strip one and leave the other and the file still talks, which is why both are removed here.

Why the pages are not re-rendered

Because a re-render would be a different document: text becomes an image, quality drops and the file grows. Metadata lives in dictionaries, so it can be deleted from the dictionaries. The trade is honesty about scope — nothing else about the file is scrubbed.

PDF metadata questions

What does a PDF give away?

Its document properties: Title, Author, Subject, Keywords, Creator and Producer, plus creation and modification dates. A browser print says “Skia/PDF”, an app export says its own name, and an AI assistant’s export sometimes says exactly that. Modern exporters also attach an XMP packet, which can carry more than the property fields.

Does this change the pages?

No. The content streams are untouched — no re-render, no flattening, text stays selectable and the page count is unchanged. Only the metadata dictionaries are deleted, which is why the file often gets slightly smaller.

Why does the result say nothing was stamped in its place?

Because that is unusual. Most PDF tools write their own producer string when they save — this site’s own PDF tools stamp “Scriptiii”, which is a fair trade for a watermark or a page-size change but wrong for a tool whose job is removal. This one saves without branding, and re-reads the written file to show that the properties are empty.

Is the document anonymous now?

No. Metadata is one channel; the text, images, fonts and any embedded file are others, and file names in the header still exist. Removal here means the properties and XMP packet are gone — it does not mean the document cannot be identified by its content.

Related tools

AI metadata checkerThe same job for images.Flatten PDFBake form fields into the page.PDF page sizeMove pages onto A4 or Letter.