Sending documents to a journalist or lawyer: clean the metadata first

Before a sensitive document leaves your Mac, check three things: the author and revision fields inside it, any tracked changes and comments it still carries, and, for photos and PDFs made from photos, the location. Word’s File > Properties and Preview’s Inspector show most of this; Word’s “Remove personal information from this file on save” setting and Preview’s Remove Location Info handle the common cases one file at a time, and Manatee’s Clean mode does it for a folder of files at once, offline. Metadata cleaning protects the document; it does not make you anonymous, and the last section is about that difference.

What a document can say about its sender

The text is what you meant to send. The rest of the file is what you did not.

A Word document (DOCX) records the name of the account that created it, the name of the last person who saved it, a revision number, total editing time in minutes, the template it was built on, sometimes a company name from the Office installation, and, if anyone used Track Changes or comments, every deleted sentence and every margin note, even when the view is set to hide them.

A PDF records Title, Author, Subject, Keywords, the application that created it (Creator), and the library that wrote it (Producer), plus creation and modification dates. A PDF made with macOS’s Print > Save as PDF sets the Author from your account’s contact card without asking.

A photo or a scan taken with a phone records the device model and, unless location was off, GPS coordinates. A PDF assembled from phone photos can carry those photos’ fields inside it.

Spreadsheets and presentations behave like Word documents.

Check before you clean

Decide what is in the file before touching it, so you know the cleaning worked.

1. For a DOCX, open it in Word and choose File > Properties. Read the Summary and Statistics tabs. Then switch Review > Tracking to show All Markup and see whether any tracked changes or comments appear.
2. For a PDF, open it in Preview and press Command-I. The General tab shows Author, Creator, and Producer.
3. For a photo, open it in Preview, press Command-I, and look for a GPS tab.
4. For anything, run mdls on it in Terminal (type mdls, drag the file in, press Return) and scan for kMDItemAuthors, kMDItemWhereFroms, and kMDItemLatitude.

A DOCX is a zip archive if you want certainty: duplicate it, rename the copy to .zip, open it, and read docProps/core.xml and docProps/app.xml in TextEdit. If word/comments.xml exists, the document has comments in it. The post on track-changes-and-comments-can-leak-cleaning-a-docx-before-you-send-it.html goes through this in detail.

Clean a Word document

1. Resolve tracked changes: Review > Accept > Accept All Changes and Stop Tracking. This is an editorial decision, not a cleaning one; read what you are accepting.
2. Delete comments: Review > Delete > Delete All Comments in Document.
3. Open Word > Preferences (or Settings) > Security and check “Remove personal information from this file on save.”
4. Choose File > Save As and save under a neutral filename. Saving as a new file, rather than over the original, gives you a clean copy and an untouched original.
5. Reopen the saved copy and look at File > Properties again. Author and last-saved-by should be empty.

Word’s privacy option clears names but can leave the revision number and editing time, and it leaves embedded images with their own metadata. For a document where that matters, see the batch method below.

Clean a PDF

What ships with a Mac is weak here. Preview shows PDF metadata and lets you edit a few fields, but there is no single “strip everything” button, and a PDF can also carry the metadata of every image placed in it. Two approaches that work:

• Flatten by reprinting. Open the PDF, choose File > Print, click the PDF button, and choose Save as PDF. The output is a fresh file without the old Creator and Producer. Fill in or empty the Title and Author fields in the save dialog, since macOS will otherwise insert your name.
• Clean it as a file. Drop it on Manatee’s Clean mode, described next, which strips the author and revision fields.

Reprinting has a cost: bookmarks and links are dropped, and a scan’s text layer can be lost. Check the result is still searchable if that matters.

There is a dedicated post on this at how-to-remove-metadata-from-a-pdf-on-a-mac.html.

Clean a folder of files at once, offline

When the material is a folder of documents, photos, and scans, cleaning each by hand is where mistakes happen: a file gets skipped, or an old copy gets sent instead of the cleaned one.

Manatee has a Clean mode that takes a batch and strips GPS, camera EXIF, and document author and revision metadata from photos, PDFs, and Word documents. It writes a new copy of every file alongside the original and never modifies the original.

1. Gather everything in one folder.
2. Open Manatee and switch to Clean.
3. Drag the folder’s contents onto the window (or onto the menu bar icon, or use Fix with Manatee from Finder’s right-click Services menu).
4. Wait for the job list to finish, then send the cleaned copies.

Two properties make it suitable for this situation specifically. Everything runs on the Mac; nothing is uploaded, there is no account, and the only network call the app makes is a license check. And its limits are stated plainly: it removes metadata, it does not redact text, it does not resolve tracked changes for you (do that in Word first), and it cannot remove what is visible in the content of a scan or screenshot. The first 24 hours are free with every feature.

Cleaning is not anonymity

Metadata cleaning removes what the file says about you. It does nothing about what the sending says about you.

• The channel. An email carries your address and your mail provider’s headers. A messaging app ties the file to your account. A cloud link ties it to the account that created the link.
• The content. A document only you could have written, a scan of a letter addressed to you, a screenshot with your name in the menu bar.
• The copy. If a document was sent to only a handful of people and yours has a unique typo, the document identifies you without any metadata at all.
• Your Mac. Extended attributes like the download URL (kMDItemWhereFroms) do not travel by email but do travel by AirDrop and inside zip archives made in Finder.

If the situation is serious, the recipient will usually have a preferred secure channel, and their guidance on that channel matters more than anything in this post. Clean the files anyway; it costs minutes and closes the easiest leak.

Questions

Is exporting to PDF from Word enough to hide tracked changes?

Only if the view is set to No Markup at export time and the changes are accepted or rejected. A PDF exported with All Markup showing prints the strikethroughs and margin comments onto the page. Resolve the changes in Word first, then export, then open the PDF and read it through.

The lawyer asked for the original file, not a PDF. Should I still clean it?

Yes. Clean a copy of the DOCX (accept changes, delete comments, remove personal information, save as new) and send that. If they need the revision history for a specific reason, they will say so; otherwise a clean copy is the expected form.

Does zipping the files protect the metadata?

No. A zip preserves everything inside each file, and a zip made in Finder adds a __MACOSX folder that carries extended attributes such as download URLs. Clean the files first, then compress if you need to.

Can Manatee remove a name that appears in the document text?

No. Clean works on metadata only. Names in the body, in headers and footers, in a signature block, or in a scanned image have to be edited or redacted by hand before the file goes out.