What is PDF Metadata? Information Dictionaries vs. XMP Packets

PDF metadata is structured information stored inside the binary code of a PDF file. There are two primary metadata storage architectures standard in modern documents:

1. Standard Document Information Dictionary (/Info)

The legacy /Info dictionary is the traditional metadata mechanism defined in the PDF specification. It consists of key-value pairs including:

• /Title: The original title of the document.
• /Author: The user account name or authoring identity of the creator.
• /Subject: Summary or topic description of the PDF.
• /Keywords: Indexing tags used by search engines and document management platforms.
• /Creator: The original application used to compose the document (e.g., Microsoft Word for Office 365).
• /Producer: The PDF creation engine (e.g., macOS Quartz PDFContext or Adobe Distiller).
• /CreationDate & /ModDate: Exact timestamps recorded down to the second when the file was created and modified.

2. Extensible Metadata Platform (XMP) Streams

Introduced by Adobe, XMP is an XML-based metadata stream embedded directly inside the PDF catalog stream. Unlike standard /Info tags, XMP packets can store extensive edit histories, Adobe Photoshop layer information, camera EXIF coordinates, and corporate workflow tracking identifiers.

To ensure complete anonymization, you must sanitize both standard /Info entries and embedded raw XMP streams. You can audit both layers simultaneously using the RiazHub PDF Privacy Stripper.

Why Stripping PDF Metadata is Essential for Security

Failing to strip metadata before publishing legal briefs, investigative journalism reports, financial earnings, or public contracts can have serious consequences:

“Unsanitized PDF metadata has repeatedly exposed whistleblowers, compromised legal confidentiality, and revealed internal corporate user account structures to cyber reconnaissance teams.”

• Whistleblower & Journalist Protection: Unsanitized author fields and creation timestamps have led to the identification of confidential sources in investigative leaks.
• Preventing Social Engineering & Phishing: Attackers analyze /Creator and /Author fields to harvest employee usernames, OS version signatures, and software build numbers for targeted spear-phishing campaigns.
• Legal & Regulatory Compliance: Law firms and financial institutions are obligated under GDPR, HIPAA, and corporate NDA agreements to scrub metadata prior to public court filings or client distribution.

100% In-Browser Privacy: How RiazHub Protects Your Confidential Data

Many online PDF tools force you to upload sensitive contracts and financial statements to remote cloud servers for processing. This creates a severe security risk, as your documents could be intercepted, logged, or indexed by third parties.

At RiazHub, we built our PDF Metadata Editor & Privacy Stripper with a strict Privacy-First Client-Side Architecture. Powered by WebAssembly and client-side JavaScript (pdf-lib), 100% of the PDF parsing, metadata extraction, editing, XMP stream stripping, and document rebuilding occurs directly inside your web browser. Your PDF files never leave your device.

Step-by-Step Guide: How to Inspect & Edit PDF Metadata Online

1. Open the RiazHub PDF Metadata Editor in any desktop or mobile browser.
2. Drag and drop your .pdf file into the secure upload zone.
3. Review the Document Diagnostic Header Card to see detected file size, page count, PDF version, and metadata health status.
4. Inspect your Privacy Exposure Risk Gauge to evaluate data leak threats (such as exposed usernames or timestamp signatures).
5. Edit any standard metadata field (Title, Author, Subject, Keywords, Producer, Creator, Creation/Mod Dates, Copyright) or click Strip All Metadata (1-Click Anonymize) for instant cleaning.
6. Add optional custom metadata key-value pairs or select professional presets like RiazHub Publishing or Clean Corporate.
7. Click Save Metadata & Download PDF to generate a clean, sanitized PDF file with automatic RiazHub branding (e.g., document_sanitized_RiazHub.pdf).