How to Strip Metadata from a PDF Before Uploading

PDFs carry more information than the visible page text. The author name from the original Word document, the application that created it, the modification timestamp, possibly the original draft comments — all of it sits in the file's metadata. When you hand the PDF to a service that does OCR, translation, or even simple conversion, that metadata travels with it.

This article walks through what's in a PDF, why it matters, and how to strip it before uploading.

What metadata lives in a PDF

A typical PDF carries three layers of metadata:

• Document Information Dictionary — the simple Title / Author / Subject / Keywords / Creator / Producer fields. Most PDF viewers expose these in the "Properties" dialog.
• XMP metadata packet — a structured XML block embedded in the PDF that can record arbitrary properties (project codes, document IDs, custom schema fields).
• Page-level metadata and resources — fonts used, image XMP for scanned pages, embedded thumbnail images.

The Document Information Dictionary is easy to inspect and easy to strip. The XMP packet is harder to see (you usually need a hex editor or a tool like EXIF Extractor), but it often carries more sensitive content.

Why this matters

If your PDF is a contract, lease, tax form, medical record, or marketing brief, the metadata can leak:

• The author's full name (often the drafter or the original creator)
• Software used (Windows version, Acrobat version)
• The internal file path on the source machine (sometimes)
• Custom XMP fields that include project numbers or client identifiers

Most PDF services don't disclose what they log from the metadata. Some use it for analytics, some for OCR training, some for compliance tracking.

How to strip it

There are three routes:

1. In Acrobat Pro: File → Properties → Description → edit each field blank. Then Tools → Sanitize Document → Remove Metadata. Slow but comprehensive.

2. In a desktop tool like pdf24 or qpdf: qpdf --remove-metadata input.pdf output.pdf. Removes XMP and the document dictionary in one shot.

3. In your browser with ToolConvoy: the Strip Image Metadata tool handles embedded image EXIF; the PDF Compress tool strips document-level metadata as a side effect.

The browser tools run entirely client-side. The file never uploads. You get back a clean PDF.

After stripping

Re-open the cleaned PDF and check "Properties" to confirm the fields are empty. If a service still received something sensitive, the cost is retroactively unrecoverable — strip before the upload, not after.

A note on metadata you actually want

Some metadata is good: the PDF title for accessibility, the language tag for screen readers, the page count for layout. Don't strip everything; strip what could leak.

The metadata fields worth keeping:

• Title (for accessibility)
• Language (en, ja, zh — for screen readers and translation)
• Producer (you can blank this; it's mostly fingerprintable)

The metadata fields worth stripping:

• Author (often your name)
• Subject (often contains project info)
• Keywords (often contains internal codes)
• Custom XMP fields (whatever your team put there)

See also

• PDF Security and Privacy Hub — the broader collection of tools for PDF metadata hygiene
• Is It Safe to Convert Files Online — the privacy argument, broader version
• Browser-Side vs Server-Side PDF Conversion — the architecture behind the choice