Document Metadata Stripper
Word, Excel, PowerPoint and PDF files carry names, companies and timings you never typed. See exactly what is in yours, then take it out — without uploading the file to do it.
Every .docx, .xlsx, .pptx and PDF has a properties block riding along inside it. Typically: the name on the template you borrowed, your employer's name, the account name of everyone who saved it, a revision counter, and the total number of minutes the document has been open for editing. It survives emailing, printing to PDF and renaming the file. Lawyers have lost cases to it and job applicants have been caught out by it. This page reads all of it on your device, shows it to you, and strips the parts you choose.
- tracked changes and deleted text still recorded in the file
- comments, and the names attached to them
- hidden rows, columns, sheets and hidden text
- the cropped-off parts of an image — cropping hides, it does not delete
- PowerPoint speaker notes
- text sitting under a black rectangle, which anyone can select and copy — for that, use Redact a PDF, which actually destroys the text
Drop a .docx, .xlsx, .pptx or .pdf below — one or a batch. Nothing is read until you drop it, and nothing leaves this tab afterwards.
Questions people ask
Does this make my document safe to publish?
No, and that is the most important thing on this page. It removes document properties — author, company, template, editing time and the rest. It does not remove tracked changes, comments, hidden rows or sheets, speaker notes, the cropped-off parts of an image, or text hidden under a black rectangle. For that last one use Redact a PDF, which deletes the text rather than covering it. A stripped file is a file with fewer properties, not a safe file.
Is my file uploaded anywhere?
No. The file is opened in your browser's memory, unzipped there, edited there and handed back to you as a download. The uploads counter in the top bar stays at 0, and your dev-tools Network tab will show the same. That matters here more than usual, because the documents people want cleaned are contracts, CVs and financial models.
What is "total editing time" and why does it matter?
Word and PowerPoint keep a running count of the minutes a document has been open for editing, and it is stored in the file. A proposal you claimed took a week can show eleven minutes; a "bespoke" report can show four, because it was a template with the names changed. It is one number and it is quietly one of the most revealing things in the file.
Will the document still open in Word after stripping?
Yes. The properties parts are kept inside the file and their values emptied, rather than deleted outright — Office is fussy about missing parts and will offer to "repair" a file that lost them. Dates and list-valued properties are the exception: an empty string is not a valid date, so those elements are removed. Nothing else in the file is rewritten; every other part is copied across with its original bytes and its original compression.
Is there a paywall, a watermark, or a file-size limit?
None of the three. Tools that do this on a server have to charge you, because they are paying for the upload, the storage and the deletion promise. This runs on your machine, so there is nothing to meter — no signup, no watermark on the output, and the only size limit is your own memory.
How this works & why it's private
A .docx, .xlsx or .pptx is a ZIP file. Rename one to .zip and you can open it yourself. Inside, docProps/core.xml holds the author, the last person to save it, the revision count and the created and modified dates; docProps/app.xml holds the application and version, the company, the manager, the template it came from and the total editing time; docProps/custom.xml holds anything your organisation added. This page opens that ZIP in your browser with JSZip, reads those three XML parts, and shows you the values.
When you strip, the XML is edited as text — the chosen elements are emptied or removed — and the archive is written back out with every other entry copied across untouched, keeping each entry's original compression method so the document content is bit-for-bit what it was. A thumbnail at docProps/thumbnail.*, which is a picture of page one and survives things people think they redacted, is deleted along with the relationship that points at it.
For PDFs, pdf-lib parses the file and the entries in the document information dictionary are deleted. Any XMP metadata stream — a second, parallel copy of the same information in XML — is unlinked from the catalogue and the object itself is removed from the file, so the text is gone from the bytes rather than merely unreferenced. If a PDF is encrypted we cannot rewrite it, and the page says so instead of handing you a broken file.
None of this needs a server, so there is not one. Open your dev-tools Network tab before you start and watch the "uploads" counter in the nav — it stays at 0 while you work.