Document privacy guide
How to Remove Metadata from a Word Document
By Laramie Williamson ·
A .docx is a ZIP archive, and some of what it stores never appears on the page. Author and last-editor names, total editing time, a company field, and template paths travel with the file when you email it. Saving a new copy does not clear any of it.
What a .docx actually stores about you
Rename a .docx to .zip and open it and the structure becomes obvious. Two small XML files hold nearly all of the identifying metadata. docProps/core.xml carries the document title, subject, the creator, the name of whoever saved it last, a revision count, and creation and modification timestamps. docProps/app.xml adds the authoring application and version, total editing time, and — often the surprising one — a Company field populated from whatever Office was licensed to.
- dc:creator and cp:lastModifiedBy commonly expose internal usernames or machine names.
- cp:revision plus TotalTime in app.xml shows roughly how long the document took.
- A stale dc:title can still say "DRAFT" or name a different client than the filename does.
<cp:coreProperties>
<dc:title>Q3 pricing — DRAFT</dc:title>
<dc:creator>j.mercer</dc:creator>
<cp:lastModifiedBy>Contractor Laptop</cp:lastModifiedBy>
<cp:revision>47</cp:revision>
<dcterms:created>2026-05-02T09:14:00Z</dcterms:created>
<dcterms:modified>2026-07-11T16:48:00Z</dcterms:modified>
</cp:coreProperties>Metadata is not the same as comments or tracked changes
This is where a metadata-only cleanup gives false confidence. Comments live in word/comments.xml, and tracked insertions and deletions are stored inline in the document body as separate markup, not in docProps. Clearing the properties files leaves every one of them in place.
If review history is the thing you are worried about, accept or reject the changes and delete the comments as explicit steps. Treat them as a different job from stripping the properties, because the file format treats them that way too.
Cropped images keep the part you cropped away
When you crop a picture in Word, the crop is a display instruction. The full original image usually stays in word/media/ at its original resolution, which means a cropped-out face, signature, or adjacent column of a screenshot is still recoverable from the file.
Word can discard that data — Picture Format, then Compress Pictures, then "Delete cropped areas of pictures" — but you have to ask for it. If a crop was the redaction, re-export the image outside Word and reinsert it instead.
Check the result rather than trusting the step
Whatever you use, re-open the cleaned copy and look at the properties again. Some fields are rewritten the moment the file is saved by an application: opening a cleaned .docx in Word and pressing save can repopulate the last-editor name with yours. That is not a failed cleanup, but it does mean the last write wins.
The FreeProTool cleanup runs in your browser, edits the properties XML in the package directly, and then rescans the output so you can see which fields are actually gone. Nothing is uploaded, which matters when the document is the thing you are trying to keep private.
Apply this to your file
When you are ready to work on the asset, use FreeProTool to remove metadata from a Word document.
Tools for this task
View and clear the properties stored inside a Word document.
Find review comments and notes, then remove selected items.
Run every check — images, links, emails, metadata, comments — in one pass.
Open the document and inspect its underlying package files.
Related guides
PDFs store two parallel sets of metadata plus annotations, form values, and attachments. Here is what is in there, and why a black box is not redaction.
A short pre-send checklist for PDF, Word, Excel, and PowerPoint files: what to look at, in what order, and which checks people usually skip.