When you share a PDF, you may be sharing far more than the visible content. Embedded inside the file are fields you never consciously filled in: your name (from your operating system's user account), the application you used and its version number, the exact date and time the document was first created and last modified, and in some cases how many revision cycles it went through. For a personal document shared with a friend, this is harmless. For a bid submission, an anonymous tip, a court filing, or a GDPR-sensitive document shared externally, these disclosures are a genuine problem.
The underappreciated complication is that PDFs carry metadata in two separate layers: the legacy Info dictionary that all readers can access, and the more detailed XMP packet that modern tools embed. Most online metadata removal tools clear only the Info dictionary — which is visible in File > Properties — and leave the XMP packet completely intact. Anyone who runs exiftool or opens Acrobat's Advanced metadata panel on your 'cleaned' PDF will see the full story. LuraPDF removes both layers, in your browser, without sending your file to any server.
Metadata removal isn't just for security researchers. Here are six real professional contexts where accidental metadata disclosure creates tangible risk.
Posting a PDF publicly — a resume, a community document, a forum attachment — embeds your full name and computer username in the file properties. Strip both Info and XMP before uploading anything you'd prefer wasn't personally attributed to you.
Court-filed PDFs with law firm metadata, attorney names, and internal document management system tags are discoverable. Clean both metadata layers before filing to control what the opposing party and public record reveal about your workflow and personnel.
Leaked or sensitive documents forwarded as PDFs carry the metadata of everyone who handled them. A journalist cleaning a document before publication — or a source before leaking — needs both Info and XMP cleared. The XMP packet often carries fields that Info-only tools miss entirely.
PDFs sent to clients, partners, or public tender processes often reveal the authoring software, internal creator name, and modification history. For a professional, brand-controlled impression, strip internal metadata before any external distribution.
Blind recruitment processes require that evaluators see only the document's content, not who created or edited it. Remove Author, Creator, and modification timestamps from scoring sheets and evaluation forms before circulating them to panel members.
Documents released under freedom-of-information requests may contain embedded metadata revealing internal author identities, system names, and edit histories beyond what redaction tools address. Full metadata scrubbing — both Info and XMP — is part of a proper pre-release sanitization workflow.
Most tools do half the job. Here's why both layers need to be cleared for metadata removal to be meaningful.
The Info dictionary is a dictionary object stored in the PDF's trailer section — the very end of the file that PDF readers parse first. It contains the standard fields: Title, Author, Subject, Keywords, Creator, Producer, CreationDate, and ModDate. LuraPDF uses pdf-lib to access this trailer dictionary directly, clears selected keys, and writes a new PDF with the dictionary updated or removed. This handles the layer that File > Properties exposes in standard PDF readers.
The XMP packet is more complex: it's a stream object embedded within the PDF that contains an XML document following the Extensible Metadata Platform specification. LuraPDF walks the PDF's object tree to locate the Metadata stream, parses the XMP XML, and removes or replaces fields across all relevant namespaces — Dublin Core (dc:creator, dc:title, dc:description), the PDF namespace (pdf:Author, pdf:Keywords), and XMPRights. Custom application namespaces added by tools like InDesign or Acrobat are also cleared. The result is a PDF where both metadata layers are clean.
| Feature | LuraPDF | Sejda / ilovepdf | Adobe Acrobat |
|---|---|---|---|
| Clears Info dictionary fields | Yes — all fields | Yes | Yes |
| Clears XMP metadata packet | Yes — all namespaces | Partial — often Info only | Yes |
| Processes file in browser (no upload) | Yes — fully local | No — server upload required | Yes — installed software |
| Free with no watermark | Yes — always free | Limited — paid for full features | Paid subscription |
Metadata removal is straightforward — but a few habits ensure nothing slips through, especially in professional or sensitive contexts.
Remove metadata before sharing, not after — once a recipient has the file with metadata, you can't un-share what they've already seen or cached
Combine with Redact PDF for full document sanitization — metadata removal cleans the file properties; redaction removes sensitive content from the visible page
Set replacement values in professional contexts — 'Anonymous' Author and 'Document Editor' Creator look intentional; blank fields can look like a scrub job to an experienced reviewer
Verify with exiftool or Acrobat after — open the cleaned PDF, run exiftool, and confirm both the Info and XMP sections show only what you intended to leave
Re-scrub after any edits — PDF editors like Acrobat, Word (Save as PDF), and LibreOffice re-stamp metadata on every save; run removal again after any revision cycle
Pair with Flatten PDF to remove annotation author metadata — flattening converts annotations to static content, stripping the per-annotation Author fields that document-level metadata removal doesn't touch
Drop your PDF, review what's embedded, and remove the fields that shouldn't be shared. Info dictionary and XMP packet — both cleared, in your browser, with nothing uploaded. No account, no watermark.