What a document carries besides its text
A Word document records who created it and who saved it last, the company its copy of Word is licensed to, when it was created, edited and printed, and the template it came from, often by its path on the firm's network. A document management system adds custom properties and variables with the matter, client and author. Comments and tracked changes keep every reviewer's name and the time of every edit. A PDF can keep every earlier revision it was saved over. None of this shows on the page, and all of it goes out with the file.
Version Story removes it before the document is sent, and reports what it found.
Two presets
| Standard | Strict | |
|---|---|---|
| Properties: author, company, dates, template, editing time, statistics, custom properties | Removed | Removed |
| People and account ids, revision ids, document ids | Removed | Removed |
| Hidden text, hidden shapes, off-slide content, unused layouts | Removed | Removed |
| Document variables, mail merge sources, custom XML, network file paths | Removed | Removed |
| Printer settings, thumbnails, picture metadata | Removed | Removed |
| A PDF's XMP, earlier revisions, JavaScript and attachments | Removed | Removed |
| Redactions marked but never applied | Applied | Applied |
| Tracked changes | Kept | Accepted |
| Comments, PDF notes, speaker notes | Kept | Removed |
| Authors' names on changes and comments | Kept | Anonymized |
| White or tiny text, hidden sheets and slides | Reported | Removed |
| Macros | Reported | Removed |
| Links to file paths and internal sites | Reported | Unlinked |
| PDF form fields, stamps and drawn signatures | Kept | Flattened into the page |
Standard is for a document you are sending as it stands: the page reads exactly as before. Strict is for a document that should arrive with no trace of how it was drafted. Tracked changes, comments and authors can each be set on their own, over either preset.
What it reports instead of removing
Some things cannot be removed without changing what the reader sees, so a clean leaves them in place and says so:
- Document numbers in a header or footer, which are visible text.
- Text under a dark box in a PDF, which looks redacted but is still in the file.
- White text over a picture or page colour, which may be legible.
- Hidden rows and columns, which formulas may point at.
- A signed PDF, whose signature no longer verifies once anything in it changes.
Every clean comes with a report of each kind of hidden data found, how many, and what was done to it.
Where to use it
- In the Version Story app: the Metadata Clean button beside Selective Compare opens a panel to drop documents into and download them clean. Or turn on Clean metadata in any download menu, and every document you download, zips included, is cleaned first with the standard preset. See Remove Metadata Before Sending a Document.
- In Claude and ChatGPT: ask for a clean copy of a document before you send it. See Claude and ChatGPT.
- Over the REST API: add
clean_metadata=standardorstrictto a file download. See Files and Metadata types.
Word, Excel, PowerPoint and PDF files are cleaned, and zips of them file by file.
Related
- Comments: how comparisons handle them
- Documents with Tracked Changes
- Security and Confidentiality
Frequently asked questions
What does a metadata clean remove?
The standard preset removes what nobody sees on the page: document properties such as author, company and dates, people and account ids, revision ids, hidden text, document variables, mail merge sources, custom XML, printer settings, image metadata, and a PDF's XMP, earlier revisions and JavaScript. The strict preset also accepts tracked changes, removes comments and speaker notes, anonymizes authors, and removes white or tiny text, hidden sheets and slides, and macros.
Does cleaning metadata change the text of the document?
Not under the standard preset: the page reads the same. The strict preset accepts tracked changes and removes comments, which is the point of choosing it, and removes text that does not show, such as white text on a white page. Anything a clean cannot remove without changing what the page shows is left in place and reported.
Can I keep comments but remove everything else?
Yes. Choose a preset, then keep, accept or reject tracked changes, keep or remove comments, and keep or anonymize authors, each on its own.
What happens to a signed PDF?
It is cleaned, and the report warns that the signature no longer verifies, since any change to a signed PDF breaks it. Clean the document before it is signed.
Can I clean a zip of documents?
Yes. Every Word, Excel, PowerPoint and PDF file inside the zip is cleaned, and the folders and file names stay as they were.
