Remove metadata from Office documents

Every Word, Excel and PowerPoint file carries a small properties block that most people never open: who created it, who last saved it, the company and manager fields from the template it started in, when it was revised and how many times — plus, often, a thumbnail image of the document. This tool clears those properties without touching the document itself, so the file still opens normally.

  • No upload
  • No account
  • No re-encoding
  • Free
Input0 bytes uploaded

Engine

File API

Runs in

Web Worker

Needs account

No

What an Office file carries

01

The properties live in `docProps/`, a handful of small XML parts beside the actual document. The table lists the ones that identify a person or an organisation.

  • Author (dc:creator)The account name of whoever created the filemedium
  • Last modified byThe account that saved it last — often a different personmedium
  • Company and managerInherited from the template or the deployment policymedium
  • Template pathOften a full network path or a share namelow
  • Title, subject, keywords, categoryFree text, sometimes a client or project codelow
  • Revision numberHow many times the file was savedlow
  • Created and modifiedLocal timestamps, including the original creation datelow
  • Total editing timeMinutes spent in the documentlow
  • Application and versione.g. Microsoft Office Word 16.0low
  • Embedded thumbnailA raster preview of the first page or slidemedium
  • Custom propertiesArbitrary name/value pairs added by templates and add-inslow

Why it matters.

02

A CV or a contract names the person who wrote it

Author and last-modified-by are the account names from the machine that saved the file. Send a CV to a recruiter and it may name a second employer; send a contract template and it may name the client whose copy you started from.

Company, manager and template leak the organisation

These fields come from the Office template rather than from anything you typed. They survive “Save As”, they survive converting to PDF in some tools, and nobody looks at them before forwarding a document outside the company.

The thumbnail can show what you removed

Office writes a small preview image into the package. If the document was edited after the preview was generated — a redacted name, a deleted slide — the old version can still be in that thumbnail. It is removed here along with the properties.

Measured against the alternative

The usual advice is “File → Info → Check for Issues → Inspect Document”, which is buried three menus deep, only exists in the desktop app, and is easy to forget. Two drags here produce the same result on any machine with a browser.

How to use it.

03
  1. Step 01

    Drop a DOCX, XLSX or PPTX file into the panel. It is read in a Web Worker on your device.

  2. Step 02

    Check the property list. Author, company and thumbnail are the fields that usually matter.

  3. Step 03

    Download the cleaned file and open it once in Office or LibreOffice to confirm it still opens without a repair prompt.

One caveat.

04

Before you rely on it

Emptying is safer than deleting

Removing the property parts outright would leave the package relationships pointing at files that no longer exist, and Word would offer to repair the document. This tool replaces each property part with a minimal valid one instead, so every relationship still resolves and the file opens normally — the difference between a cleaned file and a corrupted one.

Limits worth knowing.

06
  • Digitally signed documents are refused, not rewritten: editing the package would invalidate the signature and the file would open with a warning.
  • Comments and tracked changes are part of the document body, not of the properties. Author names inside them are left in place, because removing them would change the document.
  • The legacy binary formats (.doc, .xls, .ppt) are not supported. Save as the modern format first, then strip.
  • Custom XML parts and document variables written by third-party add-ins are not inspected.
  • Locked or password-protected packages cannot be opened, so they cannot be cleaned.

Questions people ask.

07
Does this work on XLSX and PPTX too?

Yes. All three formats use the same package structure and the same docProps parts, so the same adapter cleans them. The panel accepts all three and the report names the type it detected.

Will the document still open in Word?

It should, and the package structure is unchanged apart from the four property parts. Test the cleaned copy before deleting the original, as with any file editor — especially if the document was produced by an unusual tool rather than Office.

Why are comments and tracked changes still there?

Because they are content, not metadata. Deleting comments changes what the document says, so this tool reports them and leaves them alone. Use Word’s own review tools to accept or remove them.

Can it strip a signed document?

No, deliberately. A signature covers the package contents; changing them invalidates it. If you need the properties gone, remove the signature first, then rewrite the file.

Is the document uploaded anywhere?

No. The ZIP is opened, the property parts are replaced and the archive is rebuilt in a Web Worker on your device. Nothing is sent to a server, and there is no endpoint that accepts a file.

Files are processed on your device. Erasemetadata cannot read them.DOCX · XLSX · PPTX