How it works : Clean a PDF
A PDF does not keep its metadata in one place, it has three. The tool shows you all three, empties them, then REOPENS the file it just made to tell you what is left.
A PDF you send carries more than its pages. It says which software made it, under which account name, on which date, and often at exactly what time it was last saved. A CV exported from a word processor can carry the user name of a workstation, and a document reused from a template can carry the name of whoever made the template.
The real trouble is elsewhere: this information is written in THREE different places inside the file, and most tools clean only one. There is the classic document card, there is a second block in XML that often repeats the same things, and there is what some layout software sticks onto every page. A tool that announces “metadata removed” after emptying only the first is not lying on purpose: it does not know.
Here, all three nests are shown before anything is touched, and the tool REOPENS the file it has just produced to tell you what remains. That is exactly what the photo metadata viewer in the image corner does, and for the same reason: a cleaning promise that does not read itself back is only a promise.
How it works
- Drop your PDF in. The tool opens it and lists everything it says: the document card, the XML block, whatever is stuck on the pages, and the dates.
- Read the card. The highlighted lines are the ones that date you, name you, or name your software.
- Click “Clean and download”. The clean file arrives, and the tool tells you straight away what it removed, then what is left after reading the produced file back.
Formats accepted
PDFs only. For a photo's metadata (date, camera, GPS position), the “See and clean metadata” tool in the image corner does the same job on images.
PDF document
The page content is untouched: not the text, not the images, not the fonts, not the layout. Only the file's identity cards are removed, and the document is then REWRITTEN in full so that no earlier byte survives at the end of the file.
Its limits
What is WRITTEN in the pages stays written
This tool cleans the file's identity cards, not its content. A surname typed into the text, an address in a header, a company stamp as an image: all of that stays, and rightly so, it is your document. Deleting text inside a PDF is a different gesture, and a black rectangle placed on top does not delete it: the text is still selectable underneath. That is the classic trap of document redaction.
80 MB
The document is opened entirely in memory, and its rewritten version lives beside it. That is the honest limit of what a tab can carry without being killed by the system on a modest device.
A password-protected PDF will not open here
It has to be decrypted first, which needs its password. The “Password-protect a PDF” tool in the same corner can unlock it if you have the password.
The document identifier is reset to a shared value, not removed
The standard provides for a file identifier in the trailer, and many readers expect to find it. Removing it would make the file more fragile than useful. So it is replaced by a FIXED value, identical for every file cleaned here: two documents cleaned by this tool can no longer be told apart by that identifier, whereas the original identifier is a fingerprint of your saving session.
An electronic signature on the document does not survive
A signature seals a whole file: rewriting the file, even to clean it, invalidates it. No tool can do both, here or anywhere else. Keep the signed original aside.
Questions about this tool
Is my document sent anywhere?
No, and you can check in thirty seconds. Open your browser's inspector with the F12 key, go to the “Network” tab, clear the list, then drop your PDF in and clean it. Not a single line appears. Cut your connection and do it again: the tool behaves exactly the same.
Why three places? Is it not a single card?
No, and that is the history of the format. The PDF first had a simple card, in the file's trailer: title, author, software, dates. Then a second system arrived, in XML, richer, which often repeats the same values and sometimes adds the complete history of your saves. And some layout software sticks its own block onto every page as well. All three live in the same file, and all three need cleaning.
How do I know the cleaning really worked?
Because the tool REOPENS the file it has just made and tells you what it finds. Three endings are possible: nothing left, and it says so; something remains, and it NAMES it instead of concluding; or the rewrite failed, and it says that too. You can repeat the check yourself: drop the cleaned file into this tool, and it should answer that the file is silent.
Does this really anonymise my document?
It removes what the FILE says about you, not what the DOCUMENT says about you. If your name is typed into the text, it stays. If your signature is a pasted image, it stays. If the header carries your company's address, it stays. It is your document: a metadata tool has no business deciding to erase its content.
I put a black rectangle over a sensitive detail. Is that enough?
No, and it is the most expensive mistake in this field. A black rectangle is an object drawn ON TOP of the text: the text is still there, in the file, selectable and copyable. It can be read in seconds. To really remove information from a PDF, delete it from the source document and export again, or turn the page into an image after covering the area with an opaque fill. The image corner has a tool for masking an area of an image irreversibly.
Is the cleaned file lighter?
A little, and sometimes a lot. The XML block of a document worked on for months can weigh several kilobytes on its own, because it keeps a trace of every save. Rewriting the file in full also reclaims the space of objects that became useless. But this is not a slimming tool: for that, see “Slim down a PDF”, in the same corner.
A PDF does not keep its metadata in one place, it has three. The tool shows you all three, empties them, then REOPENS the file it just made to tell you what is left.
VerifyThe proof, step by stepThe network inspector check, explained one gesture at a time, with screenshots.
The cornerThe PDF cornerMerge, split, sign, fill, slim down. A PDF should never have to travel to a stranger to be cut in two.