.pdf ยท .pdf

What Metadata Does a PDF File Contain?

A PDF looks like a finished, flattened document, but it is really a record of the tools and people that produced it. Because saves are often appended rather than rewritten, a PDF routinely carries prior drafts, the identity of everyone who authored or signed it, and metadata that a 'scrub' left orphaned in the bytes instead of removing. File X-Ray reads all of that in your browser, so nothing is uploaded.

Scan a PDF file

What a .pdf file reveals

At the document level a scan surfaces the Author, the Creator application that made the source file (Word, Google Docs, InDesign), and the Producer, which frequently names the operating system or PDF engine that wrote it (for example macOS Quartz PDFContext or Adobe Distiller). It also reads the creation and modification dates, the XMP Document ID and Instance ID that chain every copy back to one original, and the trailer File Identifier whose first half is fixed when the document is first created and survives every later edit. Custom entries in the Info dictionary - Company and Manager from Word, or the envelope and signer IDs an e-signature platform stamps in - are pulled out too, and no PDF reader shows those.

The distinctive part of a PDF is what it keeps after you think it is gone. When a file is saved as incremental updates, each save appends new content without deleting the old, so earlier drafts are still physically in the bytes; the scan counts those retained revisions. It flags layers that are hidden by default - content invisible on screen but fully present in the file - which is a live, recoverable failed-redaction vector. And it runs a raw-byte residue scan that catches the classic scrub mistake: an ExifTool update marker plus 'orphaned' Author, Producer or Title values that a reader reports as empty while the original text still sits in the file.

A PDF also inherits metadata from whatever was placed inside it. The scan enumerates embedded XMP blocks hiding in placed photos and graphics, which commonly preserve the photographer's name, the body serial number of the camera that took the shot, an IPTC licensing contact (a working email, phone and postal address), a Staff Photographer job title, and the Lightroom or Illustrator tool that produced the asset - none of which the document's own metadata mentions. It reads the XMP 'derived from' source path and per-save edit history, and any rights, licensor or Creative Commons terms attached to the file.

Signatures and active content round it out. For each digital signature the scan reads the X.509 certificate inside the PKCS#7 blob - the signer name, organisation, serial and validity window it carries, plus the signing time recorded in its signed attributes - not just the typed name. The identity is read; the signature and chain are not verified. It surfaces the actual values typed into saved form fields, which is often the most sensitive content in the file, and outbound URLs, Launch, SubmitForm and JavaScript actions, including ones tucked inside compressed object streams that a plain byte scan would miss. Attachment filenames, embedded network paths, and encryption permission flags appear as well.

Key fields

Author, Creator & Producer
Who is named as author, the app that made the source document, and the library or OS that wrote the PDF (Producer often leaks your operating system).
Incremental Updates (retained revisions)
A count of prior document versions still present because each save was appended, not rewritten - the mechanism behind leaked drafts and failed redactions.
Hidden Layers (default-off)
Optional-content layers that are invisible when the file opens but fully present in the bytes - a live, recoverable failed-redaction vector.
Orphaned metadata & ExifTool scrub marker
Author or Producer values a reader shows as empty but that are still physically in the file, plus a %BeginExifToolUpdate marker proving a non-overwriting scrub.
File Identifier & Document/Instance ID
The trailer /ID and XMP UUIDs that link every copy and revision back to one original, even when nothing else matches.
Digital signature certificate
The signer name, organisation, serial, validity window and attested signing time read from the X.509 certificate inside the signature, not just the typed name.
Embedded asset identity
Photographer name, camera body serial, IPTC contact details and the Lightroom/Illustrator tool preserved inside placed photos and graphics.
Filled-in form field values
The data actually typed into saved form fields (names, SSNs, account numbers), flagged high when it looks identifying or lives in a password field.
Custom Info-dictionary entries
Keys like Company, Manager or an e-signature envelope/signer ID written by the authoring tool that no PDF reader displays.
Auto-run, launch & URL actions
OpenAction, /Launch external-program actions, form-submit endpoints, JavaScript and embedded URLs - resolved even inside compressed object streams.
Attachments & network paths
Embedded attachment filenames and any UNC (\\server\share) or file:// paths, which expose usernames, internal servers and case or project names.
Edit history & derived-from lineage
The per-save XMP provenance trail and the source document (and its local path) this file was exported from.

This is a selection. The full field manual documents the fields read from each format.

Questions about PDF metadata

What metadata does a PDF file contain?

At minimum an author, the application that created it (Creator), the library or OS that wrote it (Producer), creation and modification dates, and a trailer identifier. Beyond that, PDFs commonly carry XMP Document and Instance IDs, custom Info-dictionary entries like Company or an e-signature envelope ID, embedded fonts, digital-signature certificates, filled-in form values, and - because saves are appended - earlier revisions of the whole document.

Is the metadata in a PDF a privacy risk?

It can be. The Producer often reveals your operating system, the author and custom fields can name you or your employer, and absolute paths or UNC network shares embedded in the file expose your username and internal servers. The larger risk is retained content: a PDF saved as incremental updates keeps prior drafts, and layers hidden by default are invisible on screen yet fully present - the mechanism behind most failed redactions.

Why can I still recover data from a PDF that was redacted or scrubbed?

Because most PDF edits are appended, not rewritten. Drawing a black box over text does not delete the text, and running ExifTool's default command to strip metadata appends an update that blanks the fields while leaving the original bytes in place. File X-Ray flags ExifTool scrub markers, counts retained revisions, and surfaces the 'orphaned' author and producer values that a reader shows as empty.

How do I actually remove metadata from a PDF?

Use a tool that rewrites the file rather than appending to it - for example a fresh 'Save As', printing to PDF, or a sanitiser that flattens the document and overwrites the original. Simply deleting the Info fields is not enough if prior revisions remain in the bytes. This scan runs entirely in your browser, so you can check what is actually left after cleaning without uploading the file anywhere.

Can a signed PDF reveal who signed it?

Yes. Beyond the name the signer typed, File X-Ray reads the X.509 certificate inside the signature's PKCS#7 blob - the certificate's real signer name, organisation, serial number, validity window, and the attested signing time. It reads the identity but does not verify the certificate chain.

File X-Ray reads PDF files entirely in your browser - the file never leaves your device. For this format you can also download a clean copy with the metadata removed. Scan a file now.

All supported formats