How to Remove Metadata from a PDF

By the SladdPDF team · Published August 23, 2026 · Updated August 23, 2026

To remove metadata from a PDF, export a fresh copy with a tool that rewrites the file: switch on Sanitize Metadata in SladdPDF, run Sanitize Document in Adobe Acrobat, or use exiftool -all= followed by qpdf --linearize. Clearing the author field is the easy part. The same file can also carry XMP packets, attached files, comment author names, an invisible OCR text layer and earlier versions of itself that a normal save leaves behind.

What counts as metadata in a PDF?

PDF metadata is the data stored about the document rather than printed on the page — and a single file normally holds it in two independent places. The older one is the document information dictionary: Title, Author, Subject, Keywords, Creator, Producer, CreationDate and ModDate. In the specification, Creator is the application that produced the original document and Producer is the application that converted or last processed it, so the pair often exposes an entire toolchain (for example "Microsoft Word" plus a specific PDF printer build).

The second store is XMP: an XML metadata packet embedded in the file, introduced in PDF 1.4. It holds the same fields in a richer structure plus identifiers and editing history. PDF 2.0 (ISO 32000-2) deprecates most of the old info entries — everything except CreationDate and ModDate — in favour of XMP, which is exactly why a tool that only blanks the classic fields can leave a complete second copy readable.

Beyond those two stores, a PDF is a container of objects, and several of them keep data that never appears on screen:

Why does hidden PDF data matter?

Hidden data matters because it identifies people, machines and timelines that the visible document deliberately does not mention. The Author field is typically filled from the operating-system or Office account name, so an anonymous submission can arrive stamped with a real employee's login. Producer and Creator disclose software versions, which is why security researchers treat published PDFs as a reconnaissance source: in one analysis of 39,664 PDF files published by security agencies, 76% still carried the producer tool, 42% revealed the operating system, and 4% named an author, alongside leaked email addresses and internal file paths.

Institutions know the risk. The Administrative Office of the US Courts warns e-filers that a PDF can carry the author's name, the file's location on a server, the full-size version of a cropped picture and prior revisions of the text — and tells them to remove it before filing.

The canonical leak is older than the PDF era but the mechanism is identical. In October 2005, the UN's Mehlis report on the assassination of Rafik Hariri was first released as a Microsoft Word file with track changes preserved. The published text attributed the killing to "senior Lebanese and Syrian officials"; the revision history showed that an earlier draft had named individuals, including Maher al-Assad and Assef Shawkat. The visible document was correct. Its history was not — and the history is what made global headlines. Modern PDFs carry their history in different places, but the failure is the same: redaction failures are usually about what stayed in the file, not about what was drawn on the page.

What a "remove metadata" button usually misses

Most "remove metadata" features clear the document info fields and stop there, which leaves four common leak paths open.

Incremental updates keep the old version alive

PDF was designed so that edits can be appended to the end of the file instead of rewriting it. That is fast and crash-safe, but it means a saved file can contain several generations of the same object, and forensic tools identify them by looking for multiple cross-reference tables and trailers. Strip the appended section and you recover the earlier version — including, in the worst case, the content someone thought they had removed.

Attachments carry their own metadata

A PDF can embed complete files — the spreadsheet behind a table, the original Word document, an image at full resolution. Each attachment brings its own metadata, and none of it is visible on the page. This is why Adobe separates "remove metadata" from a full sanitization pass.

OCR text layers are invisible, not absent

When a scan is made searchable, the recognised text is placed over the image in an invisible rendering mode. Readers can still select, search and copy it. If you black out a name on a scanned page but keep the OCR layer, the name is one Ctrl+A away — the exact reason rasterizing the page is the safest treatment for scans.

Annotations remember who wrote them

Every comment, highlight and stamp stores an author name, normally taken from your user account. Reviewer names survive long after the comments themselves stop being relevant, which is why anonymized submissions and double-blind reviews so often fail on markup rather than on the text.

How to remove metadata from a PDF in your browser

You can remove metadata from a PDF without installing anything and without sending the file anywhere. SladdPDF runs entirely in the browser tab using pdf.js and pdf-lib: the document is read into memory on your own machine, and nothing is ever uploaded to a server. Once the page has loaded you can even disconnect from the network and finish the job offline.

1. Open the browser-based redaction tool. It loads once and then works locally.
2. Drag your PDF into the window. The file stays on your device.
3. Draw redaction boxes over anything that must not survive. Skip this step if you only want the hidden data gone.
4. Choose Secure mode and leave the Sanitize Metadata switch on — it is on by default, and it covers author, title and hidden tags.
5. Export, then reopen the result and check its document properties.

The two export modes behave differently, and the difference matters here. Secure mode renders each page to a flat image at your chosen DPI and writes a brand-new document from those images, so the original page objects — text layer, OCR layer, annotations, attachments and anything hidden beneath a redaction box — are not carried into the output at all. Fast mode keeps the original document and draws vector rectangles on it; with Sanitize Metadata on it clears title, author, subject and keywords, but the page objects stay. Treat Fast mode as a convenience for non-sensitive files, and use Secure mode whenever the hidden data is the point.

Two honest limits: the free version exports at standard resolution, and print-quality 300 DPI export requires a Pro licence key — the page count is unlimited. And rasterizing means your text stops being selectable — for a contract that still has to be searchable, a desktop tool that keeps live text is the better fit.

How to remove metadata in Adobe Acrobat

Adobe Acrobat's Sanitize Document command is the most complete one-click strip available in a mainstream PDF editor, because it targets hidden data rather than just the metadata fields. Adobe lists what it removes: metadata, embedded content and attached files, scripts, hidden layers, embedded search indexes, stored form data, comments and annotations, hidden data from previous document saves, obscured text and images, unreferenced data, links and actions.

1. Open the file in Acrobat and go to the redaction tools (Redact under Tools, in the Protection area).
2. Apply any visual redactions first and confirm them — sanitization does not black anything out for you.
3. Choose Sanitize Document to remove everything at once, or Remove Hidden Information to review and tick the categories yourself.
4. Save under a new filename, so the sanitized copy cannot be confused with the original.

Acrobat is paid desktop software, and the file is processed on your machine rather than in a browser. It is the right tool when you need to keep live, selectable text and full document structure while still clearing hidden data. Use Remove Hidden Information rather than Sanitize Document if you need to keep bookmarks or form fields.

Removing metadata from the command line

For technical users, ExifTool is the standard way to inspect and strip PDF metadata in bulk. Reading is the useful part even if you strip with something else: exiftool -a -G1 file.pdf prints every tag it can find, grouped by where it lives, which tells you instantly whether a file carries XMP as well as info-dictionary entries.

The removal command is exiftool -all= file.pdf — but on PDFs it comes with a documented catch. ExifTool edits PDFs using an incremental update, so the old metadata is not actually deleted, and the change can be undone. The fix is to rewrite the file afterwards:

• exiftool -a -G1 file.pdf — list everything that is in there.
• exiftool -all= file.pdf — clear the tags (reversible on its own).
• qpdf --linearize file.pdf clean.pdf — rewrite the file so the removal sticks.

Remember what this workflow does not do: it changes metadata, not page content. Text under a black rectangle, an OCR layer and embedded attachments all survive it untouched. For that you still need a sanitizer or a rasterizing export.

Which method removes what?

No single method covers every category, so match the tool to the risk. The table below compares the realistic options.

How do you check the metadata is really gone?

Verify the exported file, not the file you edited — and verify it the way an opponent would. Four checks catch almost everything:

1. Document properties. Reopen the export and read the description tab. Author, title, subject and keywords should be empty, and the producer should be your export tool rather than the original application.
2. Dump every tag. Run exiftool -a -G1 clean.pdf. Grouped output shows whether an XMP packet is still present after the info fields were blanked.
3. Select and copy. Press Ctrl+A, copy the page into a plain text editor and search for the names you removed. If they appear, the text layer is still there — a black box is not a redaction.
4. Open the side panels. Check the attachments and comments panels. Attached files and reviewer names are the leaks people forget, because neither is visible on the page.

If any check fails, do not patch the same file — go back to the original and re-export. Patching an already-published PDF is precisely the situation in which incremental updates preserve the very thing you are trying to remove.

This article is general guidance, not legal advice.