When you share a Word document, you share more than the words on the page. Inside the .docx file, alongside the visible text, sit properties that describe the document's history: who created it, who last edited it, which company it belongs to, and sometimes custom properties added by an organisation's templates. None of this appears in the printed page, and most of it is invisible in the normal editing view, which is why it surprises people. A document sent to a client, a journalist or a public archive can quietly reveal an internal author name, a previous editor, or a company detail that was never meant to leave the building. The DOCX Metadata Cleaner uploads a DOCX file and returns a copy with author, editor, company and related metadata removed, with the document text unchanged, entirely in your browser.
A DOCX file is not a single opaque blob. It is a ZIP archive containing XML parts: one for the document's text, one for its core properties, one for application properties, and others for styles, settings and more. The core properties part holds fields such as creator and last editor. The application properties part holds fields such as company and manager. Custom properties, when present, live in their own part. A cleaner works by editing or removing those property parts while leaving the document text part untouched. Understanding that structure explains both what the tool can promise and what it cannot: it can remove the property parts, but it cannot rewrite history embedded in the text itself, such as a name typed in a paragraph or a tracked change that records an author.
How the tool works
You upload a DOCX file, and the tool returns a cleaned copy for download. The cleaning removes creator and last editor information from the core properties, removes company and manager information from the application properties, and drops the custom properties part together with its references. The document's title is kept, and the document text is unchanged. Everything runs locally in your browser: the file is processed on your device and is never uploaded to a server. That local processing is essential for a privacy tool, because sending a sensitive document to a remote cleaner would defeat the purpose of cleaning it.
Two scope limits are stated on the tool and worth repeating. The tool works on local files only; it never touches any account, cloud drive or document service. And encrypted DOCX files are rejected, because an encrypted file cannot be read without its password, and a tool that asked for your password to clean a file would be asking for exactly the kind of access a privacy tool should not need. If your file is encrypted, decrypt a copy yourself first, clean the copy, and handle the password separately.
A worked example
Take a built-in sample document whose core properties name the creator and last editor as "Sample Author" and whose application properties name the company as "Sample Company". After cleaning, the core properties contain no creator and no last editor, and the application properties contain no company and no manager. The document text is unchanged, every paragraph reads exactly as before, and the title is kept. That before-and-after is the test to apply to any cleaner: properties gone, text identical, title preserved. If a cleaner changed the text, reordered paragraphs or dropped the title, it would have done more than cleaning, and the extra changes would need their own explanation.
Apply the same test to a real document. Before sharing, open the cleaned copy and check that the visible content matches the original: same paragraphs, same headings, same tables. Then check the properties, in whatever property viewer your word processor provides, and confirm the author, editor and company fields are empty or absent. The two checks together, text identical and properties empty, are what "cleaned" should mean. Either check alone is insufficient, because a file could keep its properties while changing its text, or change its text while keeping its properties.
Common mistakes
The most common mistake is assuming that cleaning properties cleans the document. Tracked changes, comments and hidden text can all carry author names, and a name typed in the body of the document is visible content, not metadata. Accept or remove tracked changes, delete or resolve comments, and read the document itself before sharing, because the cleaner only handles the property parts. A second mistake is cleaning the original instead of a copy. Keep the original file, clean a copy, and share the copy, so that internal records retain their full history while the shared version carries none of it.
A third mistake is trusting a file's properties display without verifying. Some viewers show cached or partial property information, so check the cleaned file in the same viewer you used for the original, and compare. A fourth mistake is forgetting that filenames, folder names and email subject lines can leak the same information the cleaner removed. A cleaned document attached to a message whose subject names the internal project has solved only part of the problem. Check the whole package you send, not just the file inside it.
Honest limitations
A metadata cleaner removes property parts; it does not guarantee anonymity. Names can survive in tracked changes, comments, headers, footers, embedded objects, document text and even in the file's revision history as stored by some systems. It works on standard DOCX files, and encrypted files are rejected rather than handled. It cannot reach into a cloud copy of the document, a version stored in a collaboration system, or an earlier copy already sent to someone else. Cleaning is one step in sharing safely, alongside reviewing the visible content, resolving changes and comments, and checking the message around the attachment. For highly sensitive material, combine cleaning with a manual review of the whole document.
Closing
Share the words, not the history. Clean a copy, confirm the text is unchanged and the properties are empty, resolve changes and comments, and check the message around the file. That routine takes a minute and prevents the quiet leak that no one notices until it is already public. For neighbouring jobs, the PDF Metadata Viewer and Editor handles the same problem in PDF files and Text and Document Comparison verifies that a cleaned copy's text matches the original, while How to Compare Two Documents and Find Changes explains document comparison in depth.