What the DOCX Metadata Cleaner does
This tool shows the hidden information inside a Word document (.docx) - the author and last editor, company, the path of the template it was made from, comments and the reviewer list, tracked changes that still hold deleted text, revision-session IDs, custom properties, document variables and the embedded thumbnail - and removes what you choose, giving you a cleaned copy and a before-and-after report.
It does what Word's Document Inspector does, without needing Word, and it proves the result by reading the cleaned file again rather than trusting the steps. The document is processed in your browser; it is never uploaded.
How to use it
- Drop a .docx file, or load the example draft report. The right-hand panel lists every category of hidden information found.
- Tick what to remove. Names, company, custom properties, comments, tracked changes, rsids, the thumbnail and the template path are ticked by default; dates, the title and keywords, and hidden text are optional because they are often wanted.
- Choose Clean document. The cleaned copy is inspected again and the table shows each item before and after.
- Download the cleaned DOCX (the original is unchanged) and, if you need a record, the CSV report.
Reading the results
Tracked changes are accepted, not rejected: inserted text stays and deleted text is removed for good. If some changes should not be kept, review them in Word first. Comments are deleted along with their anchors and the list of reviewers.
Revision-session IDs (rsids) are random numbers Word stamps on text for each editing session. They reveal nothing by themselves, but identical rsids in two files show they were edited in the same session, which can link documents.
"Remove personal information on save" is switched on in the cleaned file, so that Word does not write your name back into comments or properties the next time someone saves it.
Worked example: a council draft about to be published
The example report was written by Aisha Rahman and last edited by Marco Diaz. It records Northfield Borough Council as the company, 187 minutes of editing, revision 23, a template stored at \\nbc-fs01\templates$\arahman\ - a server and a user name - and two custom properties, including the file reference CS/2025/0412.
It also holds one comment from Marco Diaz ("Do we have the food bank figures? Legal asked before publishing."), one tracked insertion and one tracked deletion whose removed text, "Results were mixed.", is still in the file, a document variable naming the case officer, rsids and a thumbnail.
With the default options, cleaning removes all of that. The re-inspection shows no names, no comments, no revision marks, revision 1, 0 minutes of editing and Normal.dotm as the template, while the title and the created/modified dates stay because those boxes were left unticked.
What cleaning cannot remove
Anything in the visible text stays: names in the body, a signature block, an email address in a footer. So do images, which can carry their own EXIF metadata, embedded spreadsheets and other objects, and macros in .docm files. The report lists external links, embedded objects and name fields (AUTHOR, USERNAME) so you can deal with them in Word.
Limitations: what the result does not prove
- Only .docx, .docm and .dotx files are cleaned. Old .doc files and password-protected documents cannot be read here.
- Accepting tracked changes is permanent in the cleaned copy. The tool cannot reject individual changes; do that in Word first.
- Metadata inside embedded images and objects is not cleaned. Use an image metadata remover on pictures before inserting them, or replace them.
- Word may add some properties again when the cleaned file is next saved (for example the last-modified-by name of whoever saves it); the remove-personal-information setting reduces this but it is Word's behaviour, not something a file can guarantee.
Privacy: where your data goes
Everything you paste, type or drop is processed in this browser tab. It is not uploaded, logged, stored or sent to analytics. Session recording and tag-manager scripts are switched off on this page.
Standards and sources
- ECMA-376 Office Open XML File Formats - checked 19 Sep 2026
- Microsoft - Remove hidden data and personal information by inspecting documents
- NCSC - Document redaction guidance
Frequently asked questions
How do I remove the author name from a Word document?
Drop the file here with Author and last-modified-by ticked and choose Clean document. Both core properties are removed, the reviewer list is deleted with the comments, and Word's remove-personal-information setting is switched on so the name is not written back on the next save.
Does removing tracked changes accept or reject them?
It accepts them. Inserted text becomes normal text and deleted text is removed from the file entirely, which is what you want before sharing. If you want to reject some changes, do that in Word's Review tab first and then clean the file.
Can deleted text really be recovered from a Word file?
Yes, while tracked changes are in the file. The deleted words are stored in the document XML and anyone can show them with All Markup in Word. Accepting the changes, as this tool does, removes the deleted text itself.
What is an attached template path and why remove it?
Word records the full path of the template a document was created from. On company networks that path often includes a server name and a user folder, such as \\server\templates\jsmith\, which tells a reader who made the file and how your network is organised.
Is this the same as Word's Document Inspector?
It covers the same main categories - properties, comments, revisions, custom XML properties, hidden text - and adds rsids, document variables, the template path and a before-and-after check. It runs in a browser, so it also works on a computer without Word.
Is my document sent anywhere to be cleaned?
No. The .docx is unzipped, edited and zipped again in your browser, and the cleaned copy is created on your device. The original file is not modified.
Last reviewed by the A2Z.Tools team against the sources listed above.
