What Metadata Is Stored in a PDF?

A PDF stores metadata in two primary structural locations: the Document Information Dictionary (Info dict) and an Extensible Metadata Platform (XMP) XML data stream. Together, these data structures contain core document descriptors—including Title, Author, Subject, Keywords, Creator software, Producer engine, CreationDate, and ModDate—as well as rights management, font subsets, revision histories, and device identifiers.
Every Portable Document Format (PDF) file generated by word processors, desktop publishers, or virtual printer drivers carries embedded administrative and technical metadata. While visual page content consists of vector graphics, image layers, and glyph positioning instructions, PDF metadata resides outside the visual canvas within dedicated dictionary tables and XML streams. Standardized under ISO 32000-1 (PDF 1.7) and ISO 32000-2 (PDF 2.0), this attribute payload enables operating systems, search engines, and enterprise indexers to categorize and manage files. However, because metadata is written automatically during creation, PDF documents frequently persist confidential authorship handles, local workstation file paths, editing durations, and software toolchain signatures without the sender’s explicit awareness.
Key Definitions: PDF Metadata, Document Info Dict, XMP Stream, and ISO Standards
To analyze the internal structure of a PDF file, technical definitions must be established for the standards and containers governing embedded metadata payloads.
- PDF Metadata: Non-visual attribute data stored within a PDF binary container. PDF metadata defines document properties, authorship identity, creation software, security parameters, and modification histories independently of visual page layouts.
- Document Information Dictionary (Info dict): The legacy key-value dictionary table defined in early PDF specifications (PDF 1.0–1.7). Positioned near the trailer cross-reference table, the Info dict maps predefined keys (such as
/Authoror/CreationDate) to literal string objects. - XMP (Extensible Metadata Platform) Data Stream: An ISO-standardized framework (ISO 16684-1) created by Adobe that embeds an XML metadata payload formatted according to W3C Resource Description Framework (RDF) syntax. Referenced in the PDF Document Catalog (
/Rootcatalog object), the XMP stream supports custom namespaces, nested properties, and granular editing histories. - PDF 1.7 / ISO 32000-1 and PDF 2.0 / ISO 32000-2: International specifications governing PDF file construction. ISO 32000-1 allows dual metadata storage in both the Info dict and XMP stream, whereas ISO 32000-2 formally deprecates legacy Info dict fields in favor of synchronized XMP packets.
Information Dictionary (Info Dict) Fields: Standard Keys and Syntax
The Document Information Dictionary represents the original metadata store defined in ISO 32000-1 Section 14.3.3. Attributes shown under basic viewer properties are retrieved from this dictionary. The specification defines eight standard fields:
- Title (
/Title): A text string specifying the actual document title, often derived from manuscript headers rather than the physical file name. - Author (
/Author): A text string identifying the document creator, automatically populated with the operating system username, login profile, or software registration handle. - Subject (
/Subject): A text string summarizing the document subject matter or topic abstract. - Keywords (
/Keywords): A string of comma-separated or space-separated search keywords for enterprise indexers. - Creator (
/Creator): A text string identifying the original authoring software (e.g.,Microsoft® Word for Microsoft 365,Adobe InDesign 19.0, orGoogle Docs). - Producer (
/Producer): A text string naming the underlying rendering library, virtual printer, or conversion tool that compiled the final PDF file (e.g.,Acrobat Distiller 24.0,Skia/PDF m122, orQuartz PDFContext). - CreationDate (
/CreationDate): An ASN.1-formatted date string recording when the PDF was compiled, formatted as(D:YYYYMMDDHHmmSSOHH'mm')with UTC time zone offsets (e.g.,D:20260918083000-04'00'). - ModDate (
/ModDate): A date string recording the timestamp when the PDF was last modified or resaved.
Within the underlying binary stream, the Info dict is serialized as an indirect object table:
12 0 obj
<<
/Title (Quarterly Audit & Risk Assessment 2026)
/Author (Elizabeth Vance)
/Subject (Enterprise Security Compliance)
/Keywords (audit, compliance, security, 2026)
/Creator (Microsoft Word for Office 365)
/Producer (Acrobat Distiller 24.0)
/CreationDate (D:20260918083000-04'00')
/ModDate (D:20260918114522-04'00')
>>
endobj
XMP Metadata Data Stream: XML Schemas, Namespaces, and Data Payload
Modern PDF documents rely on XMP streams to store structured metadata payloads. Standardized under ISO 16684-1, XMP embeds XML data packets directly inside a stream object linked from the Document Catalog (/Catalog dictionary under /Metadata).
Wrapped inside XML processing instructions (<?xpacket begin="..." id="W5M0MpCehiHzreSzNTczkc9d"?>), XMP packets can be read by external asset management systems without full PDF binary parsing. XMP organizes attributes into standardized XML schema namespaces:
- Dublin Core Schema (
dc:namespacehttp://purl.org/dc/elements/1.1/): Contains universal descriptors includingdc:title,dc:creator(author array),dc:description,dc:subject,dc:publisher,dc:rights, anddc:format(application/pdf). - Adobe PDF Schema (
pdf:namespacehttp://ns.adobe.com/pdf/1.3/): Encapsulates PDF-specific properties such aspdf:Keywords,pdf:PDFVersion,pdf:Producer, andpdf:Trapped. - Photoshop & Basic Schemas (
photoshop:,xmp:): Stores creation dates (xmp:CreateDate), modification timestamps (xmp:ModifyDate), and editing tool states. - Media Management Schema (
xmpMM:namespacehttp://ns.adobe.com/xap/1.0/mm/): Tracks file lineage through GUIDs (xmpMM:DocumentID,xmpMM:InstanceID) and anxmpMM:Historyresource array logging every save, tool invocation, and export step. - Rights Management Schema (
xmpRights:): Encodes copyright status, legal statements, and licensing terms.
Compliant PDF 1.7 and PDF 2.0 parsers give authoritative precedence to XMP streams when values conflict with legacy Info dict entries.
Hidden & Invisible Metadata Risks in PDF Files
PDF documents contain non-visual data structures that introduce privacy exposures beyond standard fields:
- Incremental Update Trailers & Revision History: PDF specifications support append-only saving, where edits are appended to the file end alongside a new cross-reference (
xref) table. Deleting text or pages in standard editors leaves original unedited bytes and earlier metadata states intact in prior byte offsets. - Embedded Font Subsets & PostScript Metrics: Embedded fonts contain descriptors (
/FontDescriptor) with system font names and glyph metrics that can reveal local workstation font paths. - Annotations, Markup Layers, and Deleted Comments: Reviewer notes, highlights, and sticky notes reside in the
/Annotsarray. Even when hidden on screen, annotation objects retain author handles, edit timestamps, and full comment strings. - Workstation Usernames and Local File Paths: Creator strings and OLE links frequently retain absolute workstation file paths (e.g.,
C:\Users\jdoe\Documents\Confidential\Tender.docx), exposing network directory structures and user identity handles. - Embedded File Attachments & Printer Identifiers: PDF containers can embed source files within
/EmbeddedFilesdictionaries or store application data in/PieceInfotables, alongside device printer identifiers and trapped rendering flags.
How to View and Inspect PDF Metadata in Your Browser
Inspecting metadata before distributing files is critical to prevent accidental data leaks. The safest method uses client-side, browser-based tools that process document structures in memory without remote server uploads.
You can view every hidden attribute inside any document using the PDF metadata extractor. When a file is loaded, client-side JavaScript reads the PDF binary buffer (Uint8Array), locates the trailer /Info dictionary, and parses the /Metadata object stream to render the XMP RDF/XML payload. Because all processing occurs locally within your browser sandbox, sensitive reports and legal documents never leave your local device memory.
How PDF Metadata Differs from Office (DOCX/XLSX) Metadata
Comparing PDF metadata to Microsoft Office metadata highlights structural differences across file formats:
- Container Architecture: Office files (DOCX, XLSX, PPTX) follow the Open Packaging Conventions (OPC)—a zipped container of structured XML files (
docProps/core.xml,docProps/app.xml,docProps/custom.xml). PDF files store metadata in binary dictionaries and XMP streams within a continuous binary stream. - Revision History Tracking: Word documents record Revision Save Identifiers (RSIDs) inside
word/document.xmlto track editing sessions. PDF files track history through incremental trailer revisions and XMP media management history arrays (xmpMM:History). - Conversion Persistence: Exporting an Office file to PDF extracts metadata from
docProps/core.xmlto populate the PDF/Infodict and XMP payload, while stamping the conversion engine name into the/Producerfield.
To inspect hidden attributes in office documents, use the Office metadata extractor.
Frequently Asked Questions
What metadata is stored inside a standard PDF document?
A standard PDF contains Document Information Dictionary fields (Title, Author, Subject, Keywords, Creator software, Producer engine, CreationDate, and ModDate) and an XMP XML stream holding Dublin Core descriptors, software licenses, font metrics, and modification histories.
Where is metadata located within the PDF file structure?
PDF metadata is stored in two locations: the Document Information Dictionary (located near the PDF trailer xref cross-reference block) and the XMP metadata stream (an XML stream object referenced under the /Metadata key in the Document Catalog).
What is the difference between the PDF Info dictionary and an XMP metadata stream?
The Info dictionary is a legacy key-value string container defined in PDF 1.0–1.7. The XMP metadata stream is an ISO 16684-1 standard RDF/XML data packet introduced in PDF 1.4 supporting complex nested schemas, custom namespaces, and media management histories.
Can PDF metadata reveal the identity of the author or software user?
Yes. The /Author field frequently embeds operating system usernames or software registration handles, while /Creator and /Producer fields record software builds, rendering engines, and local workstation file paths.
What are PDF incremental updates and how do they expose deleted metadata?
Incremental updates append modifications to the end of a file without rewriting existing binary objects. Editing sensitive text or metadata standardly leaves original content fully readable in earlier byte offsets of the binary file.
Is PDF metadata visible when printing or viewing the document page?
No. Metadata resides within non-visual structural containers in the PDF file code. It is never drawn onto the visual page canvas during screen rendering or physical printing unless explicitly converted into visual text header elements.
How can I extract and inspect PDF metadata without uploading my document to a server?
You can inspect PDF metadata securely using an in-browser tool like the PDF metadata extractor, which parses binary byte arrays in local device RAM without sending files over remote web networks.
Related Reading & Tools
- How to Remove Hidden Metadata Before Sharing a PDF
- Browser PDF Metadata Extractor
- Office Metadata Extractor
- What Is File Metadata? Types, Examples, and How to View It
Sources & References
- ISO 32000-1:2008: Document management — Portable document format — Part 1: PDF 1.7. International Organization for Standardization. Technical specifications for document information dictionaries and metadata stream structures.
- ISO 32000-2:2020: Document management — Portable document format — Part 2: PDF 2.0. International Organization for Standardization. Deprecation of legacy Info dictionary entries and XMP synchronization rules.
- ISO 16684-1:2019: Graphic technology — Extensible Metadata Platform (XMP) specification — Part 1: Data model, serialization and core properties. Standardized RDF/XML metadata data model and namespace definitions.
- Adobe Technical Note #5144: PDF Reference, Version 1.7 (Sixth Edition). Adobe Systems Incorporated. Implementation guidelines for document catalog metadata, incremental updates, and dictionary objects.
Last updated: 18 September 2026.