PDF files are not just pages of text and graphics. Hidden inside every PDF is a layer of metadata: who created the file, which software was used, when it was last modified, and sometimes even information you never intended to include. If you are sending a PDF to a client, including it in a bid proposal, or submitting your resume, that metadata can reveal more about you than you realize. This guide shows exactly which fields are exposed, how to inspect them, and three practical ways to strip them out, plus the traps that make clean PDFs harder than they look.
A standard PDF contains an information dictionary with fields like:
On Windows, PDFs generated by Word, Excel, or even a browser Save as PDF will carry the author's computer name, full username, or software version. On macOS, Preview and Quick Look stamp the file with the logged-in Apple ID name or Mac account name. If you have ever sent a PDF and then realized the file said Created by JohnDoe on PC-JH0921 in Microsoft Word, you know this is not hypothetical.
Before removing anything, confirm what is actually in the file. Three quick methods cover most users.
pdfinfo ships with poppler and is available on Linux, macOS via Homebrew, and Windows via poppler-utils binaries. It prints the information dictionary cleanly:
pdfinfo document.pdf
You will see lines like:
Title: Q3 Report
Author: Jane Smith
Creator: Microsoft Word 2019
Producer: Microsoft Print to PDF
CreationDate: D:20260901103012Z
ModDate: D:20261108141200Z
This is the fastest audit. Pipe it to grep if you want to check only specific fields:
pdfinfo document.pdf | grep -E 'Author|Creator|Producer'
The inspector lists Title, Author, Subject, Keywords, and Creation Date directly. You can also right-click the inspector rows and edit them inline, a convenient way to blank out the Author field before sharing. The limitation: Preview does not expose /Creator or /Producer, and it will not remove embedded EXIF data from images inside the PDF.
Sites that parse PDF headers and display the info dictionary are useful for a fast visual check, especially on mobile. Upload the file, and the tool will list every key/value pair in the metadata block. Use this when you are not on your own machine and do not have poppler installed.
Treat any online tool as untrusted for confidential documents. For client contracts, bids, or anything under an NDA, stick to local pdfinfo or Preview.
Several free web services will take a PDF, strip the info dictionary, and return a clean file.
Workflow:
Pros: No software install, works on any device, fastest option for a one-off file.
Cons: Your file leaves your machine. Do not use this for legal documents, financial filings, or anything under an NDA. Also, some free tools only clear the top-level info dictionary and leave XMP packets, embedded image EXIF, or comment annotations untouched. Always re-audit.
Preview is built into every Mac and is the lowest-friction option for Mac users.
Workflow:
Pros: No install, no file leaves the machine, works offline.
Cons: You cannot blank /Creator or /Producer this way. Those fields are written by the producing application and Preview does not expose them in the inspector. If the PDF was generated by Word or Chrome on a Mac, /Creator and /Producer will survive a Preview save. For those fields you need Method 3 or a scripted solution.
For a complete scrub, command-line tools give you control over every field.
Option A — qpdf, recommended and cross-platform:
brew install qpdf on macOS, or apt install qpdf on Linux.
qpdf --replace-input --empty --new-input document.pdf
The --empty flag clears the entire info dictionary, and --replace-input writes back to the same file. This is the single most reliable one-liner for stripping /Info on any platform.
Option B — Ghostscript, if qpdf is unavailable:
gs -sDEVICE=pdfwrite -dNOPAUSE -dBATCH -dQUIET -dPDFINFO={/Title () /Author () /Creator () /Producer ()} -sOutputFile=cleaned.pdf document.pdf
Option C — Python with pikepdf, for batch jobs:
import pikepdf, pathlib
for f in pathlib.Path('pdfs').glob('*.pdf'):
pdf = pikepdf.open(f)
del pdf.Root[pikepdf.Name('/Info')]
pdf.save(str(f))
This removes the entire /Info tree. For batch processing a folder of PDFs before a submission deadline, pikepdf is the fastest and most reliable option.
Stripping the info dictionary is necessary but not sufficient. PDFs hide data in places most cleanup tools miss.
Every JPEG or PNG embedded in a PDF carries its own EXIF block: GPS coordinates, camera model, creation timestamp, and sometimes the photographer's name. pdfinfo will not show these. To find them:
mutool extract doc.pdf images/
exiftool -all= images/*.jpg
mutool comes from the MuPDF tools package, and exiftool is a standalone binary available on all platforms. In practice, the easiest path is to regenerate the PDF from source files that have had their EXIF stripped beforehand. On macOS, use Photo, then Tools, then Remove Location Data. On Windows, right-click an image, select Properties, go to the Details tab, and choose Remove personal info.
If your document uses a non-standard or company-branded font, the PDF will embed a subset of that font and record the font name in the font dictionary. pdfinfo does not list font names, but you can check:
qpdf --json input.pdf | grep -i font
If a proprietary or identifiable font name leaks, replace it with a generic font before exporting, or run Ghostscript pdfwrite, which re-embeds fonts under generic names. That usually resolves the issue.
PDFs can carry annotation objects such as sticky notes, highlights, and form fields, and even embedded file attachments. These are invisible in most viewers unless you open the annotation panel. Before sharing, flatten the document:
gs -dNOPAUSE -dBATCH -dQUIET -dPDFINFO={/Comments ()} -sDEVICE=pdfwrite -dEmbedStandardFonts -sOutputFile=flat.pdf doc.pdf
Or in Preview on macOS 13 or later, use File, Export as PDF, with the flatten option enabled.
Modern PDFs often carry an XMP block inside the file, separate from the classic /Info dictionary. Tools that only clear /Info will leave XMP intact. qpdf --empty handles both. If you use a simpler tool, verify the XMP packet is gone with:
qpdf --json doc.pdf | python3 -c "import sys,json; d=json.load(sys.stdin); print(d.get('metadata',{}))"
If it still contains author or software strings, run qpdf with --empty to clear it.
Run through this list before hitting Send on any PDF that leaves your machine:
PDF metadata is not a minor detail. It is a fingerprint of who made the file, with what software, on what machine, and when. A 30-second pdfinfo audit catches the obvious fields. A full scrub with qpdf plus EXIF, font, and annotation cleanup handles the rest. Build the pre-send checklist into your workflow, and the next time someone opens your document, they will see only the content you intended to share.
What metadata is typically hidden in PDF files?
Common hidden fields include the author name, title, creation date, software version, and company details. Some files contain keywords or subject lines that may be visible only in properties. Metadata can also hold GPS coordinates from photos or email addresses if scanned from digital sources. Review these fields before sharing sensitive documents publicly.
How do I view PDF metadata in my browser?
Most modern browsers like Chrome or Firefox allow you to view metadata without plugins. Open the PDF, right-click, and select 'Properties' or check the document info section in the sidebar. Alternatively, use online viewers that display a dedicated metadata panel. This lets you inspect author, creation software, and modification timestamps instantly.
Can I remove metadata without re-saving the PDF?
No, metadata is embedded in the file structure. To remove it, you must process the file through a tool that strips these fields. Our browser tool processes the PDF client-side, removing author, title, and software tags, then downloads a clean version. The original file remains unchanged unless you save the new version, ensuring your local copy stays intact.
Is it safe to upload sensitive PDFs to web tools?
Security depends on implementation. Look for tools that process files entirely in your browser using JavaScript, never uploading data to servers. Verify that no network requests occur during processing. For highly confidential documents, use offline software or ensure the service explicitly states client-side processing to prevent data leakage or unauthorized storage of your personal information.