Remove metadata from Word documents

DOCX properties and package contents

A DOCX package can carry core, application and custom properties, comment authors and revision history. Relationships, custom XML, media, thumbnails or embedded objects can add further information. Inspect the package findings alongside the text in Word; a document property view is not a full account of those parts.

• Core/extended/custom properties and authorship
• Legacy/modern comments and tracked revisions
• Relationships, custom XML, thumbnails, media and embedded objects

DOCX text changes during processing

The supported path removes assessed document properties and privacy-bearing package metadata, then checks the output independently. It also removes comments and accepts tracked insertions or move-to text while discarding tracked deletions or move-from text. Review that resulting text before disclosure rather than treating revision removal as a neutral property change.

Retained DOCX text and package structure

The output retains supported document text after the tracked-change rules are applied, with required package declarations and relationships reconciled to the remaining parts. That does not preserve the review history or every embedded object. Compare the processed document with the disclosure you intend and keep the working original where required.

• Supported document text after tracked insertions and move-to text are accepted, and tracked deletions and move-from text are discarded.
• Required package content types and relationships reconciled to the remaining parts.

DOCX scope and review requirements

This page covers DOCX, without extending the claim to DOC or macro-enabled formats. The current path does not promise fully selective retention of chosen fields. Comments, deleted revisions, field instructions or embedded objects can disappear, so review the report and compare the resulting text before relying on the processed document.

• The supported scope is DOCX; DOC and DOCM do not inherit it.
• The current workflow does not promise fully selective retention of chosen metadata fields.
• Comments, deleted revisions, field instructions and embedded objects can disappear.
• Review the resulting text and content-removal report before using the processed document.