Manatee

Track changes and comments can leak: cleaning a DOCX before you send it

· 7 min read

Switching Word’s view to No Markup hides tracked changes and comments on your screen; it does not remove them from the file. Every deleted sentence, every margin note, and the name of whoever wrote them stays inside the DOCX until you accept or reject the changes and delete the comments. Before sending, use Review > Accept All Changes and Stop Tracking, Review > Delete All Comments in Document, turn on “Remove personal information from this file on save” in Word’s Security preferences, and save as a new file. Then verify, because the verification is the part people skip.

What a DOCX actually contains

A DOCX is a zip archive of XML files. You can see this yourself: duplicate the document in Finder, change the copy’s extension from .docx to .zip, and double-click it. Inside:

  • word/document.xml is the text. Tracked insertions are wrapped in <w:ins> tags and deletions in <w:del> tags, each with an author name and a timestamp. The deleted text is right there, in full.
  • word/comments.xml exists if the document has or has had comments. Each comment carries the author, their initials, a date, and the text.
  • docProps/core.xml holds the creator, the last person to modify it, the revision number, and the created and modified dates.
  • docProps/app.xml holds total editing time in minutes, the template the document was based on, sometimes the company name from the Office installation, and word and page counts.
  • word/media/ holds embedded images, each with whatever EXIF they had when they were pasted in.

Word’s view settings (All Markup, Simple Markup, No Markup) only decide which of this to draw on screen. The recipient’s Word opens the same zip and can choose any view they like.

How the leak usually happens

Three patterns account for most of the cases.

The negotiated document. A contract or proposal goes back and forth with Track Changes on. The final version is sent with the view set to No Markup. The other side switches to All Markup and reads every position you retreated from, with dates.

The internal review. Colleagues comment on a draft (“can we really claim this?”, “legal hasn’t approved this section”). The comments are resolved in conversation but never deleted from the file. Resolved comments are still in comments.xml.

The reused template. A new document is made by opening an old one and replacing the text. The core properties still name the original author, the revision number is in the dozens, and the editing time is hours, all for a “fresh” two-page letter.

None of these require the recipient to do anything clever. They only require opening the file.

Clean the document in Word

Work on a copy, keep the original, and do the steps in this order.

  1. Open the document and set Review > Tracking to All Markup, so you can see what you are about to resolve.
  2. Read through the changes. Choose Review > Accept > Accept All Changes and Stop Tracking. If some changes should be rejected instead, handle those individually first; this is an editorial step as much as a cleaning one.
  3. Choose Review > Delete > Delete All Comments in Document. Resolved comments count; delete them too.
  4. Open Word > Preferences (called Settings in newer versions) > Security. Under Privacy options, check “Remove personal information from this file on save.”
  5. Choose File > Properties and clear anything left in the Summary tab (Title, Subject, Author, Company, Comments).
  6. Choose File > Save As, give the file a neutral name, and save it as a new DOCX.
  7. Close the document. Reopen the new file and check File > Properties again; the author and last-saved-by fields should be blank.

Two caveats with Word’s privacy option. It clears names, but in some versions it leaves the revision number and total editing time in app.xml. And it does nothing to the EXIF inside embedded images, so a pasted phone photo can still carry a location inside the document.

There is a shorter post focused on the author field alone at how-to-remove-the-author-name-and-edit-history-from-a-word-document.html.

Strip the remaining metadata with Manatee

For the fields Word leaves behind, or for a folder of documents, Manatee’s Clean mode strips the document author and revision metadata and writes a new copy alongside the original.

  1. Resolve the tracked changes and delete the comments in Word first, as above. Deciding which edits stand is editorial work and belongs in Word.
  2. Open Manatee and switch to Clean.
  3. Drop the DOCX files onto the window, onto the menu bar icon, or select them in Finder, right-click, and choose Fix with Manatee under Services.
  4. Wait for the job list to finish and send the cleaned copies.

Everything runs on the Mac; nothing is uploaded and there is no account. The original file is never modified. The boundaries are worth stating: Manatee cleans metadata, it does not decide your tracked changes for you, it does not rewrite text in the body, and it does not remove a name that is typed into the letterhead. Those are yours to handle in Word before the file goes in.

Verify the file is clean

This is the step that turns “I think I cleaned it” into “I know I did.”

  1. Duplicate the cleaned DOCX in Finder and rename the duplicate to end in .zip.
  2. Double-click it to expand, and open the folder.
  3. Confirm there is no word/comments.xml. If it exists and is essentially empty, the comments were deleted; if it has <w:comment> entries, they were not.
  4. Open word/document.xml in TextEdit and use Edit > Find to search for w:ins and w:del. There should be no matches.
  5. Open docProps/core.xml and confirm dc:creator and cp:lastModifiedBy are empty.
  6. Open docProps/app.xml and look at TotalTime, Company, and Template.

Alternatively, open the cleaned file in Word, set All Markup, and scroll through. That catches the tracked-changes case but not the properties.

If the recipient only needs to read it

When the other side does not need to edit, a PDF is the safer format, with one condition: export it with the view set to No Markup after accepting the changes and deleting the comments. A PDF exported with All Markup showing prints the strikethroughs and comment balloons onto the page, which is a worse leak than the DOCX because it looks deliberate.

After exporting, open the PDF in Preview, press Command-I, and check the Author field. macOS and Word both tend to fill it in for you. The post on how-to-remove-metadata-from-a-pdf-on-a-mac.html covers cleaning that side.

Questions

If I turn off Track Changes, does that remove the changes?

No. Turning tracking off stops recording new changes. The existing insertions and deletions remain in the file until each is accepted or rejected.

Are resolved comments removed from the file?

No. Resolving marks a comment as done and dims it on screen. The comment, its author, and its text stay in comments.xml until it is deleted.

Does copying the text into a new blank document fix everything?

It removes tracked changes and comments, since only the final text is pasted. It does not help with the properties, because the new document gets your name as author and the current date, and pasted images keep their EXIF. It is a reasonable first step, not a complete one.

Does Pages or another word processor have the same problem?

Any editor that supports change tracking stores the history in the file. The menus differ, but the principle is the same: accept or reject everything, delete comments, then check the exported file rather than the screen.