Remove metadata from a PowerPoint deck

What we remove

• Core properties — creator, last modified by, revision count, timestamps, title, subject, description, category, content status and keywords.
• Extended properties — company, manager, template name, editing time and the hyperlink base URL.
• Custom properties and speaker-note comment threads.
• The embedded thumbnail, a preview render of the first slide.

Slides, media and layouts are preserved.

What we deliberately keep

• Every slide, layout and theme in the package — slideLayouts, slideMasters and theme.xml are untouched.
• Embedded media and fonts, copied through exactly as they were.
• Speaker notes text itself — only comment threads attached to them, and their authors, are removed.

What a real dump looks like

A .pptx package follows the same ZIP-of-XML-parts structure as Word and Excel. docProps/core.xml and app.xml carry creator, company and template name — the template name is specific to PowerPoint and often reveals which internal or client-facing template a deck started from, which can matter more than the author's name if the deck is being shared outside the organisation. Comment threads live in ppt/comments/, one part per slide, each entry carrying an author and timestamp independent of the slide's visible content.

The package format and slide/comment schema are part of ECMA-376, the same Office Open XML standard behind .docx and .xlsx — one spec, one packaging convention, applied consistently across all three formats.

What we don't reach: embedded objects

Paste a live Excel chart into a slide rather than a static picture, and PowerPoint doesn't just store the rendered image — it embeds the actual source file as a nested package, typically under ppt/embeddings/. That nested file is a complete .xlsx or .docx in its own right, with its own core.xml and app.xml, carrying its own creator and company fields independent of the deck's. A pass over the outer package's XML parts doesn't open that nested archive, so an embedded object's metadata travels with the deck untouched. If a slide has a live embedded chart or table rather than a picture, it's worth running that source file through its own cleaning pass, or converting it to an image if you don't need it to stay linked and editable.

What survives common editors

Exporting to PDF for distribution strips slide comments entirely, since PDF has no equivalent structure, but the deck's core properties often survive into the PDF's own Info dictionary through the conversion. "Duplicate" or "Save As" in PowerPoint keeps the original core properties unless you explicitly clear them via Document Inspector. The template-name leak is the one worth watching for decks specifically: a deck built from an internal template, then genericised for an external audience by deleting a logo or two, usually still names the original template file in its properties — a giveaway that the "custom" deck wasn't built from scratch.

Hidden slides aren't metadata

A slide marked hidden still presents in the file when you open it — "hidden" only means PowerPoint skips it during Present mode, and every viewer that opens the raw file can see it anyway. We treat that the same way we treat named ranges in a spreadsheet: it's content the author chose to include, not a property describing the file, so it's outside what a metadata pass touches. If a deck has a hidden slide with numbers you don't want the recipient to see, deleting the slide is the only fix — not a metadata tool, this one included.

Cleaning something else?

Sharing the deck as a PDF export too? That copy carries its own metadata and needs a separate pass. See the document formats hub for how the same properties risk shows up across formats. The same properties and comment risk apply to Word documents and Excel spreadsheets from the same source.