PDF Archival Standard

PDF/A-3: Embed Any File Inside an Archival PDF

PDF/A-3 (ISO 19005-3) is the archival PDF standard that allows any file type — XML, CSV, CAD, Excel, JSON — to be embedded inside a permanently self-contained PDF. The outer PDF is fully archival-compliant. The inner attachment carries machine-readable source data. One file serves both humans and automated systems.

Quick Answer

PDF/A-1 and PDF/A-2 created archival PDFs but made a trade-off: no embedded files (or only other PDF/A files). PDF/A-3 breaks this restriction. You can embed any file — an XML invoice, a CAD drawing, a live Excel spreadsheet — directly inside a PDF/A-compliant wrapper. The printed version is readable by humans forever. The embedded XML is readable by accounting software, ERP systems, and government portals. One document, two audiences, one file. This is how European e-invoicing (ZUGFeRD / Factur-X) works.

What Is PDF/A-3?

PDF/A-3 is part of the ISO 19005 family of archival PDF standards. Published in 2012, it builds directly on PDF/A-2 (ISO 19005-2) with a single, crucial extension: the embedded file attachment restriction is removed.

All three generations of PDF/A share the same core archival requirements for the outer PDF wrapper:

  • All fonts must be fully embedded — no system font dependencies that could change over time.
  • All colour spaces must be device-independent — colours render identically on any display or printer, now and in the future.
  • No encryption — archival documents must be readable without passwords or keys that may be lost.
  • No external dependencies — all content needed to render the page is contained within the file.
  • No JavaScript or executable actions — which could behave differently across different PDF viewer versions.

What PDF/A-3 uniquely adds is the ability to attach any file — using the PDF Associated Files mechanism — to the document or to individual objects within it, while still declaring the outer PDF as fully conformant to the archival standard.

📌

The archival guarantee applies to the outer PDF only. PDF/A-3 does not impose archival requirements on the embedded attachments. A ZUGFeRD XML attachment is preserved because it is embedded inside the archival PDF — not because XML itself is archival-certified.

PDF/A-1 vs. PDF/A-2 vs. PDF/A-3

FeaturePDF/A-1 (2005)PDF/A-2 (2011)PDF/A-3 (2012)
Transparency✗ Not allowed✓ Allowed✓ Allowed
JPEG 2000 (JPX)✗ Not allowed✓ Allowed✓ Allowed
Optional content (layers)✗ Not allowed✓ Allowed✓ Allowed
Embedded file attachments✗ None⚠ PDF/A files only✓ Any file type
ZUGFeRD / Factur-X e-invoicing✓ Standard basis
Conformance levelsA, BA, B, UA, B, U

Real-World Examples

🧾 Finance Scenario

ZUGFeRD E-Invoice: One File, Two Readers

A German manufacturing company issues 10,000 invoices per month. Each invoice is a PDF/A-3 file. The human-readable page shows a beautifully formatted invoice with logo, line items, and payment terms. Embedded inside the same file is a Factur-X XML document containing every line item as structured EN 16931 data. When the PDF arrives at the customer's ERP system, the accounts payable software reads the XML automatically — zero manual data entry. When the customer's accountant opens the file, they see the formatted PDF. One file, one email, two audiences.

🏗️ Engineering Scenario

Construction Firm: Archived Drawings With Live Source Files

A construction firm archives all project documentation as PDF/A-3. Each drawing PDF embeds the original AutoCAD .dwg file as an attached file. In 2040, when a building requires renovation, the facility manager opens the 20-year-old PDF — the drawing is clearly visible as guaranteed by the archival standard. They click the attachment and extract the original CAD file directly from the PDF, with all editable layers and geometry intact. No separate file server required.

🔬 Research Scenario

Scientific Publisher: Paper With Embedded Dataset

A physics journal requires all papers to be submitted as PDF/A-3 with the underlying research dataset embedded as a CSV or HDF5 attachment. Reviewers read the formatted paper. Data analysts download the embedded dataset directly from the PDF without hunting for supplementary files in a separate repository. In fifty years, an AI research assistant can still read the structured data from a file that remains perfectly preserved in the archival wrapper.

Why PDF/A-3 Matters

🗄️

True Self-Contained Archives

Source data, formatted output, and all supporting files live inside one archival container. No broken links, no missing attachments, no separate file management needed.

🤖

Machine-Readable + Human-Readable

The PDF serves humans visually and machines programmatically — the foundation of modern e-invoicing (ZUGFeRD, Factur-X, EN 16931) across Europe and globally.

♾️

Long-Term Preservation

The outer PDF is guaranteed readable in any ISO 19005 conformant viewer decades into the future — regardless of software changes, OS updates, or vendor decisions.

🔗

No External Dependencies

Fonts, colour profiles, ICC data — all embedded. The file is completely self-contained and never relies on external servers, cloud services, or operating system resources.

📋

Compliance-Ready

PDF/A-3 is the mandated format for e-invoicing in France (Factur-X), Germany (ZUGFeRD), and is aligned with EU e-invoicing directive EN 16931 for B2G procurement.

🏛️

Government Archives

National archives and government document management systems increasingly require PDF/A-3 for digital filings that carry embedded structured data alongside human-readable content.

PDF/A-3 Conformance Levels

LevelWhat It RequiresUse Case
PDF/A-3a (Accessible)Full structural tagging — heading hierarchy, tables, lists, alt text for images, reading order. Screen reader compliant.Government documents, public sector, regulated industries
PDF/A-3b (Basic)Visual appearance preserved — fonts embedded, colours device-independent. No structural tagging required.Most commercial e-invoicing, ZUGFeRD, general archiving
PDF/A-3u (Unicode)All text Unicode-mapped for copy/paste and search. Not fully tagged for screen readers.Searchable archives that don't require full accessibility

Common Mistakes to Avoid

  • Assuming the embedded attachment is also archival. PDF/A-3 guarantees the outer PDF wrapper is archival. The embedded file (e.g., a proprietary CAD format) carries no such guarantee. If long-term readability of the attachment matters, choose an open format for the embedded content.
  • Encrypting a PDF/A-3 document. Encryption is explicitly prohibited by the PDF/A standard — an encrypted file is not conformant. Never password-protect a PDF/A-3 file. Use access rights management at the system level instead.
  • Not validating conformance after creation. Many PDF tools claim to export PDF/A-3 but produce non-conformant files. Always validate with a dedicated PDF/A validator (e.g., veraPDF, a free open-source ISO validator) before archiving or distributing.
  • Using JavaScript or form actions. PDF/A strictly prohibits JavaScript, launch actions, and other dynamic behaviours. Forms must use static AcroForm fields — XFA forms are not permitted in PDF/A-3.
  • Choosing PDF/A-3 when you need PDF/A-4. PDF/A-4 (published 2020, based on PDF 2.0) adds further improvements. For new projects starting in 2024+, evaluate whether PDF/A-4 better meets your needs — particularly if you work with PDF 2.0 tools.

Frequently Asked Questions

  • PDF/A-3 (ISO 19005-3, 2012) is an archival PDF standard that extends PDF/A-2 by allowing any file type — XML, CAD, CSV, Excel — to be embedded as an attachment inside a fully archival-compliant PDF. Used for e-invoicing (ZUGFeRD/Factur-X) and document archives that carry source data alongside the human-readable version.

  • PDF/A-1: Strict baseline — no embedded files. PDF/A-2: Adds transparency, JPEG 2000, layers, allows embedding other PDF/A files. PDF/A-3: Identical to PDF/A-2 except any file type can be embedded. The visual page archival requirements are identical across all three.

  • ZUGFeRD / Factur-X is a European e-invoicing standard that uses PDF/A-3 as its container. Each invoice is a PDF/A-3 file with a human-readable invoice page plus an embedded Factur-X XML (EN 16931 compliant) containing structured machine-readable invoice data — one file readable by both humans and automated accounting systems.

  • Use PDF/A-3 when you need to embed source data files (XML invoices, CSV datasets, CAD files) alongside the archival PDF — and need one self-contained file to serve both human readers and automated systems. Use PDF/A-2 for pure archival PDFs with no machine-readable attachment requirement.

  • Yes — the outer PDF/A-3 wrapper fully meets the archival standard (fonts embedded, device-independent colours, no external dependencies). The embedded attachments are preserved inside the file but are not themselves subject to archival requirements — their long-term readability depends on their own format's openness.

  • Level A: Full structural tagging for screen reader accessibility. Level B: Visual appearance preserved — no tagging required. Most ZUGFeRD files use Level B. Level U: Text Unicode-mapped for search but not fully tagged. Validate conformance with veraPDF before using in production.

Create Archival-Grade PDFs — Free

PDFlyst helps you convert, compress, and manage PDFs that meet professional archival standards.

Open PDF Editor — Free