How to Remove Metadata from a Word Document (2026 Guide)

Why Word Metadata Has Caused Real-World Disasters

Before we get into the how-to, it's worth understanding what's actually at stake. These aren't theoretical privacy risks — they're documented incidents that made international headlines.

In February 2003, the British government published what became known as the "dodgy dossier" — a document presented to justify the invasion of Iraq. Within hours of publication, a Cambridge academic named Glen Rangwala noticed something odd. He renamed the Word file to a .zip archive, opened the XML, and found revision metadata showing the document had been copied almost word-for-word from a graduate student's thesis by Ibrahim al-Marashi. The revision history, author names, and edit trail were all intact inside the file. The government had never stripped the metadata before releasing it publicly. The resulting scandal, reported by the BBC and later confirmed in parliamentary proceedings, became one of the most famous metadata leaks in history.

The second case is less famous but arguably more damaging legally. In 2004, the SCO Group filed a breach-of-contract lawsuit in Oakland County, Michigan — against DaimlerChrysler. Except the Word document they filed still contained tracked-changes metadata showing the original defendant in the lawsuit was Bank of America. Someone had simply done a find-and-replace on the company name without stripping the revision history. Reporters at Neowin and Groklaw published the exposed metadata, and SCO was left explaining why they had apparently planned to sue a completely different company. The "wrong defendant" was visible to anyone who opened the document's XML.

What's Actually Hidden Inside a .docx File

Most people think of a Word document as a single file. It isn't. Rename any .docx file to .zip and you'll find a folder structure inside — and three XML files that store metadata separately from your document text.

docProps/core.xml is the most personally sensitive. It stores the original author (the name from your Office account when the document was first created), the last person to modify the file, creation date, last modification date, and a revision counter showing exactly how many times the document has been saved.

docProps/app.xml stores the application version used to create or edit the file, the company name from your Microsoft 365 licence, total editing time accumulated across all sessions (in minutes), word count, page count, and the template the document is based on. That editing time field is particularly revealing — it can show that a document supposedly written in an hour actually accumulated 14 hours of editing, suggesting significant revision that wasn't disclosed.

docProps/custom.xml holds any custom document properties your organisation has defined — things like project codes, department names, or internal classification labels.

Beyond those three property files, the document body XML itself (word/document.xml) contains tracked changes as inline markup. Accepted changes don't disappear from this markup — they're marked as accepted, but the original deleted text and the inserted replacement are both still present in the file. The same applies to comments: even after you delete them in Word's interface, their XML elements may linger in the file.

There's also a fifth category that surprises most people: embedded objects. If your Word document contains a spreadsheet, PDF, or image that was inserted using Insert > Object (rather than just copy-pasted), those embedded objects carry their own metadata separately. Document Inspector doesn't always catch embedded objects in all versions of Word — something we'll come back to.

Method 1: How to Remove Metadata from a Word Document Using Document Inspector

Microsoft built Document Inspector specifically to surface and remove hidden data. It's the right first step for any document you're preparing to share. Here's the exact process across the current versions of Word.

On Windows (Microsoft 365 / Word 2019/2021):

Open the document. Click File in the top-left ribbon, then click Info in the left panel. Under "Inspect Document", click "Check for Issues" and select "Inspect Document" from the dropdown. Word may prompt you to save before inspecting — do it. In the Document Inspector dialog, all categories are checked by default. Leave them all checked and click Inspect.

Document Inspector will scan for: Comments and Revisions (all tracked changes and comments), Document Properties and Personal Information (author name, company, last modified by, etc.), Headers and Footers, Hidden Text, Invisible Content, Embedded Documents, and Macros. Next to each category where data was found, you'll see a Remove All button. Click it for every category. Then click Close and save the document.

On Mac (Microsoft 365 for Mac):

The path is slightly different on macOS. Open the document, then go to Tools in the menu bar > Protect Document. In the Privacy section at the bottom, check "Remove personal information from this file on save" and click OK. Save the document. This handles the core.xml and app.xml fields. For a more thorough inspection including tracked changes, use Review > Track Changes > Accept All Changes, then follow up with the Windows version if possible or use MetaClean's Office tool.

After running Document Inspector and removing everything it found, open the document in a fresh session and run it again. This second pass catches anything the first pass may have missed — particularly in complex documents with embedded objects or nested content.

What Document Inspector Misses

Document Inspector is useful but not complete. This is the part most guides skip, and it's important.

The first gap: revision counter. Even after stripping author names and tracked changes, the revision number in core.xml can reveal how many times the document was saved. A document with 87 saves that's presented as a "first draft" is telling a story you may not want told. Document Inspector removes personal information but doesn't reset the save counter in all versions.

The second gap: embedded objects. If your document contains an embedded spreadsheet, the spreadsheet carries its own metadata — its own author, its own revision history, its own edit time. Document Inspector's "Embedded Documents" check handles this in most cases, but in some older versions of Word it doesn't dig into nested objects. If you're sharing a document with embedded Excel data and privacy matters, open the embedded object and clean it separately.

The third gap: custom XML parts. Some enterprise Word templates and add-ins write data into custom XML parts that Document Inspector doesn't surface at all. These can contain document classification labels, workflow status fields, or even user account identifiers from your organisation's systems.

The fourth gap: the template link. Word stores a reference to the template the document was based on — and that reference can include a UNC network path like \\internal-server\templates\LegalDraft.dotx. That path exposes your internal network structure to anyone outside your organisation who opens the file.

Method 2: Manual XML Cleanup (Advanced)

For anyone who wants to see exactly what's in their document — or who needs to clean fields that Document Inspector doesn't touch — the manual XML approach is straightforward. You don't need any special software beyond a text editor.

First, make a copy of the .docx file. Then rename the copy from document.docx to document.zip. On Windows, right-click and Extract All; on Mac, double-click to open. You'll see a folder structure including a docProps folder and a word folder.

Open docProps/core.xml in any text editor (Notepad, TextEdit, or VS Code). You'll see fields like <dc:creator>Jane Smith</dc:creator> (original author), <cp:lastModifiedBy>John Doe</cp:lastModifiedBy> (last editor), and <cp:revision>47</cp:revision> (save count). You can edit the values directly — replace names with empty strings or generic text, or remove the elements entirely.

Open docProps/app.xml and look for <Company>Acme Corporation</Company> and <TotalTime>840</TotalTime> (editing time in minutes). Edit or clear these as needed.

After editing, re-zip the folder contents (select all files and folders inside, not the folder itself), and rename the .zip back to .docx. Open it in Word to verify it still works correctly.

This approach gives you precise control over exactly what stays and what goes. But it's manual, it's easy to make mistakes, and it doesn't scale if you're cleaning multiple documents. For most people, Document Inspector plus a browser-based tool is the right combination.

Method 3: Windows File Properties (Quick but Limited)

Windows offers a faster path that works without opening Word at all. Right-click the .docx file in File Explorer, select Properties, and click the Details tab. At the bottom you'll see "Remove Properties and Personal Information." Click it, choose "Remove the following properties from this file," check everything, and click OK.

This removes the properties visible in the Details pane — author, company, last saved by, creation date, and so on. It's fast and doesn't require Word to be installed. But it has a hard limitation: it only touches the properties visible in the Details tab. It doesn't remove tracked changes, comments, or the revision history stored in the document body XML. Think of it as skimming the surface rather than a thorough clean.

Use Windows File Properties as a quick sanity check or for documents where you're confident tracked changes aren't an issue — say, a document you drafted entirely yourself in a single session. For any document that went through collaboration, review cycles, or multiple authors, you need Document Inspector or a dedicated tool.

Method 4: MetaClean's Office Metadata Remover (Browser-Based)

For documents where you need comprehensive cleaning without the risk of missing something, MetaClean's Office metadata remover handles .docx files entirely in your browser. Nothing is uploaded to a server — the file is read locally by your browser's File API, cleaned in memory, and offered back as a download.

The tool processes all three property XML files (core.xml, app.xml, and custom.xml), strips tracked change markup from the document body, removes comment annotations, and clears the template reference path. It works on Windows, Mac, Linux, iOS, and Android — anything with a modern browser — without installing software or creating an account.

Drop your .docx file into the MetaClean DOCX tool, click clean, and download. The process takes a few seconds even for large documents. If you're cleaning multiple documents, the batch mode handles them all at once.

In our experience processing Office documents with MetaClean, the most commonly found data is the company name from the Office licence — something Document Inspector doesn't always surface prominently. The total editing time field is a close second. Both fields regularly surprise people who assumed their documents were clean.

How to Verify the Cleanup Worked

Don't just trust the process — verify it. After cleaning, rename the .docx to .zip and open it. Look at core.xml and app.xml directly. The creator and lastModifiedBy fields should be empty or contain generic text. The Company field in app.xml should be blank. The TotalTime field can remain (it's less sensitive) but you may want to zero it out.

In Word/document.xml, look for <w:ins and <w:del elements — these are tracked insertions and deletions. If they're still present, tracked changes weren't fully removed. Also look for <w:comment elements in word/comments.xml. If that file is empty or absent, comments are gone.

For the template path, open word/settings.xml and search for attachedTemplate. If the element is absent or the value is empty, you're clean. If it contains a \\server\path style reference, that's a network path that should be removed.

This manual verification step takes about two minutes and gives you certainty rather than assumption. For any document heading to a client, court, or media contact, that two minutes is worth it.

When Metadata Removal Is Non-Negotiable

There are specific situations where sending a Word document without stripping metadata first is genuinely risky.

Legal filings and contracts. The SCO case is the canonical example, but it's not isolated. Lawyers have accidentally revealed their strategy, their client's initial position, and even privileged communications through tracked-changes metadata in documents filed with courts or sent to opposing counsel. The American Bar Association's technology committee has published guidance specifically about this risk. Many firms now have policies requiring Document Inspector to be run before any document leaves the building.

Press releases and public statements. The Blair dossier showed what happens when a government document goes out with its edit trail intact. The same applies to any organisation releasing a public document — every revision, every name that touched the file, every deleted sentence is potentially recoverable. For more on the business risks of document metadata, that risk extends to PDFs and Office files equally.

Job applications and freelance proposals. Your CV and cover letter were likely drafted in Word. They carry your full name (expected), but also your company name from your Office licence (which might be your current employer's name — revealing where you currently work to a recruiter who doesn't know yet), and potentially the names of anyone who reviewed your draft.

Documents shared with competitors or vendors. Pricing proposals, RFP responses, and vendor assessments often go through multiple internal reviewers. The internal discussion in tracked changes — the "this number is too low, push back to X" comment that someone forgot to delete — can be visible to the recipient. This is a documented source of leaked negotiating positions.

Résumés uploaded to LinkedIn. If you upload a Word document as your LinkedIn résumé, the metadata travels with the file. Our guide on LinkedIn document metadata covers exactly what's at risk when you attach Office files to your professional profile.

What About Excel and PowerPoint?

The same risks apply. Excel (.xlsx) and PowerPoint (.pptx) files use the same ZIP-plus-XML structure as Word. They have their own core.xml and app.xml with the same author and company fields. Excel adds another layer: formula cell history and named ranges that may contain sensitive internal labels. PowerPoint often carries embedded notes in the speaker notes pane — notes that weren't meant for the audience but travel with the file when it's shared.

Document Inspector in Excel and PowerPoint works the same way as in Word: File > Info > Check for Issues > Inspect Document. Run it before sharing any Office file. MetaClean's Office tool handles .xlsx and .pptx alongside .docx, so you can clean all three formats in the same session.

For PDF documents — which have their own separate metadata risks — our complete PDF metadata removal guide covers the full process including XMP metadata, author fields, and creation software traces. And if you're specifically trying to remove your author name from a PDF before sharing, our guide on removing author names from PDFs goes through every method.

Prevention: Reducing Metadata at the Source

Cleaning documents before sharing is the right habit. But you can also reduce how much metadata accumulates in the first place.

In Word's Privacy settings (File > Options > Trust Center > Trust Center Settings > Privacy Options on Windows), check "Remove personal information from file properties on save." This tells Word to strip author names from the core.xml on every save — so the metadata never accumulates in the first place. It doesn't help with tracked changes or comments, but it means you're not constantly fighting the same author-name problem.

Turn off Track Changes when you don't need it. The metadata only accumulates when tracking is on. If you're drafting a document yourself and don't need revision history, keep tracking off. You can always enable it before sending for review, then accept everything and strip the tracked-change markup before the final version goes out.

Use a generic Office account name for documents you know will be shared publicly. Some organisations create a generic "Communications" or "Marketing" user account for drafting documents that will go to press — this way the author field carries a department name rather than an individual's name.

Consider converting to PDF as the final sharing format. PDFs have their own metadata risks (covered in our PDF guides), but converting from Word to PDF using Word's built-in export strips tracked changes, comments, and revision history — they simply don't exist in the PDF format. The PDF's metadata fields (author, creator, creation date) still need to be cleaned, but you've eliminated the DOCX-specific risks.

Frequently Asked Questions

Does accepting all tracked changes remove the revision history from the file?

No. Clicking "Accept All Changes" in Word removes the visual markup from your document view, but the underlying XML in the .docx file still contains the tracked-change elements. Word marks them as accepted, but the deleted text and original author names remain in the file structure. You need to run Document Inspector (or a dedicated tool like MetaClean) to actually purge that data from the XML.

Can I remove Word metadata on a Mac without the full Windows Document Inspector?

Partially. On macOS, go to Tools > Protect Document and enable "Remove personal information from this file on save" — this clears the author fields from core.xml. For tracked changes, use Review > Accept All Changes. But Mac Word's built-in inspector is less thorough than the Windows version, so for sensitive documents, either use MetaClean's browser-based Office tool or verify the cleanup manually by inspecting the XML inside the renamed .zip archive.

What's the fastest way to check if a Word document has hidden metadata?

Rename the .docx file to .docx.zip, extract it, and open docProps/core.xml in any text editor. You'll see author names, dates, and revision counts in plain text within seconds. For tracked changes, look for <w:ins> and <w:del> elements in word/document.xml. This takes about two minutes and shows you exactly what's there before you share anything.

Does Document Inspector remove everything?

Not quite. Document Inspector handles author names, tracked changes, comments, headers and footers, and most hidden text. But it has documented gaps: it may not catch metadata inside deeply nested embedded objects, it doesn't always clear the revision counter, and it doesn't consistently remove the template network path stored in word/settings.xml. A manual XML check or a dedicated tool covers those gaps.

Will removing metadata break or corrupt my Word document?

Removing metadata from the property XML files (core.xml, app.xml) is completely safe — those files are separate from your document content. Removing tracked changes and comments via Document Inspector is also safe, though it permanently removes the revision history. The only scenario where cleanup can cause issues is removing certain custom XML parts that a Word add-in or enterprise system relies on — if your document uses specialised templates, test the cleaned version before distributing it widely.

Do the same metadata risks apply to Google Docs?

Google Docs handles metadata differently. There's no embedded XML metadata in the same sense — author and revision information is stored in Google's servers, not embedded in the file. However, when you export a Google Doc as .docx, the export can embed your Google account name as the author. If you export to share externally, run the downloaded .docx through Document Inspector or MetaClean before sending it on.