RedactDrop › Guides › Remove metadata from a PDF before sending it
Remove metadata from a PDF before sending it
Published 27 September 2026 · by Tekiba
Short answer: a PDF can carry an author's name, the software that made it, dates, a second copy of its title in XMP, attached files, bookmarks, comments and even an earlier version of itself. To take all of that out, open RedactDrop's redaction page, drop the PDF on it, leave the words empty and press Redact and check. What each page itself draws comes out as it was. Form fields and comments are extras too: they sit on top of the page, so a filled-in form or a visible comment is left out unless you choose to keep its look — look at every page of the result.
What a PDF carries besides its pages
The pages are only part of a PDF file. Next to them, the format has places for information about the document and for things that travel with it. Most viewers show little of this unless you go looking, which is why it gets sent without anyone meaning to.
To show what that looks like, here is what a one-page staffing report made up for this article carries. The values are what pdf.js, the PDF reader used by Firefox, reads from the file; the section numbers are those of the PDF specification, ISO 32000-1.
| Where | What the example file holds |
|---|---|
| Document information dictionary (14.3.3) | Title “Q3 staffing review”, Author “Morgan Example”, Subject “Internal: salaries and restructuring”, Keywords “confidential, salaries, Northfield office”, Creator “Example Writer 16.2”, Producer “Example PDF Library 4.1”, created 14 September 2026, modified 26 September 2026 |
| XMP metadata stream (14.3.2) | Title “DRAFT Q3 staffing review for Jordan Sample”, creator “Morgan Example”, the same two software names |
| File identifiers (14.4) | Two 16-byte document IDs |
| Embedded files (7.11.4) | An attachment called salary-notes.txt |
| Document outline (12.3.3) | A bookmark: “Appendix: names to be moved” |
| Annotations (12.5.6.4) | A comment: “Check these numbers with Jordan before sending” |
| Incremental updates (7.5.6) | Two saved generations of the file; the first still holds the old title |
Two things in that table are worth a second look. The XMP title is not the title in the document information: whoever finished the report changed one and not the other, so the draft title — with a colleague's name in it — is still in the file. And the file has been saved twice by appending the changes to its end, which PDF allows; the first version is still in there, byte for byte, underneath the second. A viewer shows the latest version, but anyone reading the file itself can find the earlier one.
How much of it a viewer shows you
The document properties dialog of most viewers shows the document information: title, author, subject, keywords, the creating and producing software, and dates. Attachments, bookmarks and comments have panels of their own, if you open them. XMP metadata and earlier saved versions are usually not shown at all. So “I looked at the properties and cleared the author” leaves a good deal behind.
Take it out with RedactDrop
RedactDrop builds a new PDF from what is needed to draw the pages, and leaves everything else out. You do not have to remove any words to use it that way.
- Open the redaction page and drop the PDF on it. It is read in your browser and not uploaded.
- Leave Words to remove empty and draw no boxes.
- Press Redact and check. The new file is opened again and checked before it is offered.
Reading the new file with pdf.js again finds no document information, no XMP, no attachments, no bookmarks, no comments and a single generation; the old title is not anywhere in the file's bytes. pdf.js still reports an ID for it, but that is a fingerprint pdf.js works out from the file when the file has none of its own.
What the new file keeps
These are left out, whether or not they hold anything sensitive:
- the document information dictionary (title, author, dates), XMP metadata and the document ID;
- ActualText and alternate text, and the structure tree (the tags screen readers and search use);
- annotations (comments), form fields, links, bookmarks and attachments;
- scripts, and earlier generations of the file left by incremental updates.
Form fields and comments are among them, so a filled-in form loses its values and a visible comment disappears unless you keep their look. Three of them can be kept if you tick them under “What the new file keeps” on the redaction page: web links and links to pages; bookmarks, whose titles are then checked for any words you are removing; and the look of form fields and comments, drawn into the page, where the file stores how they look. What you keep of the look does not include the fields' data.
Leaving out the structure tree has a cost worth knowing: a tagged PDF made for screen readers loses its tags. If accessibility matters for the copy you send, keep the original for that purpose and send the rebuilt copy where only the pages matter.
Other ways to do it
Adobe Acrobat Pro has a step for this that Adobe calls sanitizing, which removes hidden information such as metadata, comments and attachments (Adobe: Sanitize PDFs). Many PDF editors let you clear the document properties by hand, which removes the document information but not necessarily XMP, attachments or earlier versions.
How to check a file yourself
- Open its document properties and look at every field, including the creating and producing software.
- Open the attachments, bookmarks and comments panels.
- If you have a text editor that can open a PDF as text, search it for a name or a phrase from the old title: older generations and uncompressed XMP show up that way.
Take the extras out of a PDF without uploading it. The free version handles one PDF of up to three pages. Open the redaction page.
Questions
Does removing metadata change how the pages look?
What each page itself draws is rebuilt from the same drawing instructions and looks the same. Form fields and comments are not part of that: they sit on top of the page, so by default a filled-in value or a visible comment is left out along with the metadata. Tick the option to keep the look of form fields and comments, which draws them into the page where the file stores how they look, and check every page of the result.
Is clearing the author in the document properties enough?
Often not. The same name can be in the XMP metadata, in a comment, in an attachment or in an earlier saved version of the file, and the properties dialog does not show those.
Do I have to type any words?
No. Leave the words empty and draw no boxes, and RedactDrop only rebuilds the file without its metadata and other extras.
Is the PDF uploaded?
No. It is read and rebuilt in your browser tab; nothing from it is sent to this site.