How to Remove Metadata From a PDF

A PDF stores its metadata in two places, and tools that clean one often ignore the other.

• The document information dictionary. The old mechanism, a simple list of key/value pairs: Title, Author, Subject, Keywords, Creator, Producer, CreationDate, and ModDate. Author is usually your operating-system account name, filled in without asking you. Creator names the application you authored in; Producer names whatever wrote the actual PDF bytes.
• The XMP metadata stream. An XML packet embedded in the file, and the place modern software prefers. It can hold everything the info dictionary holds, plus edit history, a document identifier that survives across saved revisions, and — when the file came out of a photo application — EXIF properties copied from the source image, GPS among them.

Clearing the Author field in a PDF editor typically rewrites the info dictionary and leaves the XMP packet as it was, so the name you deleted is still in the file — just not where the Properties dialog looks for it. Any tool you trust must handle both.

Yes, though never in the field you would think to check. The document information dictionary has no latitude or longitude entry — there is nowhere in it for coordinates to live. Location reaches a PDF by three other routes.

• Through the XMP packet. Export a geotagged photograph to PDF from an image editor and the EXIF properties, including exif:GPSLatitude and exif:GPSLongitude, can be copied into the document's XMP alongside everything else. Nothing in the Properties dialog will show them to you.
• Through the pictures inside it.A PDF embeds each image as its own object. Depending on how the producing application built the file, that object may still contain the photo's original EXIF block, coordinates and camera serial number intact. Stripping the document's metadata does not touch metadata belonging to the images within it.
• Because the PDF is a map. The geospatial features standardised in PDF 2.0 (ISO 32000-2), and shipped earlier as an Adobe extension to PDF 1.7, let a page carry a viewport with a geographic coordinate system attached, mapping points on the page to real-world latitude and longitude. This is how the USGS US Topo maps work. It is a feature, not a leak — but if you cropped a screenshot out of one, the georeferencing may have come along.

To see coordinates that came in with an embedded image, ExifTool has to be told to descend into the file:

# -ee walks embedded images and documents, not just the PDF itself
exiftool -ee -a -G1 document.pdf

# Just the GPS: prints nothing if there is none
exiftool -ee -gps:all document.pdf

If that finds GPS, the surest fix is to strip the photographs before they go into the document — how location gets into a photo in the first place covers where those coordinates come from and how to stop your camera writing them.

Windows has no built-in PDF metadata remover. The Remove Properties and Personal Informationlink in a file's Properties dialog only works for formats Windows has a writable property handler for — JPEG and TIFF do; PDF normally does not, and the link is greyed out. Our Windows EXIF removal guide covers that dialog for the formats where it does work.

ExifTool and qpdf, the reliable route

# Install ExifTool (winget ships with Windows 11 and current Windows 10)
winget install -e --id OliverBetz.ExifTool

# Clear the info dictionary and the XMP packet
exiftool -all= document.pdf

# Rewrite the file so the cleared metadata is not recoverable
qpdf --linearize document.pdf clean.pdf

Reopen PowerShell after installing so it picks up the new PATH. qpdf is distributed as a Windows build on its releases page; unzip it and either add the bin folder to your PATH or call qpdf.exe by full path. Delete the document.pdf_original backup ExifTool leaves behind — it still has everything.

A whole folder at once

exiftool -all= -overwrite_original -r -ext pdf C:\Docs

Get-ChildItem C:\Docs -Recurse -Filter *.pdf | ForEach-Object {
  qpdf --linearize --replace-input $_.FullName
}

--replace-input rewrites each PDF in place. -ext pdf keeps ExifTool from wandering into any images sharing the folder.

Adobe Acrobat Pro, if you have it

File → Properties → Description lets you clear Title, Author, Subject, and Keywords, and Additional Metadata reaches the XMP packet behind them. But the tool you want is Tools → Redact → Sanitize Document, which removes metadata along with embedded scripts, hidden layers, attachments, and deleted-but-still-present content, then saves a rebuilt file. Sanitizing is not reversible and it is not in the free Acrobat Reader.

Before it becomes a PDF

If you still have the Word document, clean it there and the export starts clean: File → Info → Check for Issues → Inspect Document, then remove Document Properties and Personal Information. Save as PDF afterwards. This also catches tracked changes and comments, which survive a PDF export and which no metadata stripper will find, because by then they are page content.

Preview shows you a PDF's metadata under Tools → Show Inspectorand will not let you edit any of it. Neither will Finder's Get Info. macOS has no built-in editor for this, the way it has none for photo EXIF beyond the Photos app's location toggle.

ExifTool and qpdf, via Homebrew

brew install exiftool qpdf

exiftool -all= document.pdf
qpdf --linearize document.pdf clean.pdf

# Or a folder of them, rewritten in place
exiftool -all= -overwrite_original -r -ext pdf ~/Documents/Contracts
find ~/Documents/Contracts -name '*.pdf' \
  -exec qpdf --linearize --replace-input {} \;

The same two passes, for the same reason: ExifTool clears, qpdf rewrites.

Print to PDF, when you cannot install anything

Open the document, press ⌘ + P, and choose Save as PDFfrom the PDF menu. macOS regenerates the document through its own rendering pipeline, so the file you get is a new one, described by macOS's metadata rather than the original's.

You pay for it. Bookmarks, links, form fields, and any text layer added by OCR may not survive, and the print job can put the original filename into the new document's Title. It is a reasonable move on a locked-down machine, and one to check afterwards rather than trust.

Search for this and you will find a dozen sites offering to strip your PDF in the browser. Read what they mean by that. Most upload the document to a server, process it there, and give you a download link — which means the file you were trying to make anonymous has been handed, in full, to a company you had not heard of ten seconds earlier, along with your IP address and the time you sent it.

If the metadata bothers you enough to remove it, the document probably should not be uploaded to remove it. That is the whole argument, and it does not have an answer. Use one of the local methods above for anything you would not post publicly. If you do use an online service, prefer one that says explicitly that processing happens in your browser, check that the page keeps working after you disconnect from the network, and assume anything you upload is retained.

The same reasoning drives our own tool: it runs entirely in your browser and nothing is transmitted. It reads and strips images, not PDFs — which makes it the right tool for the step before this one, cleaning photographs so the document you build from them starts with nothing to remove.