What is PDF/UA?

PDF/UA stands for “Universal Accessibility” and is the international standard for accessible PDF documents, set out in the ISO 14289. It precisely defines which technical properties a PDF file must meet so that assistive technologies - such as screen readers, Braille displays or reading functions - can reliably capture and reproduce the content.

What is crucial is that PDF/UA does not regulate what a document looks like, but rather how it is internally structured. Two PDFs can appear identical on the screen - only the document with a complete structural level (tags, reading order, alt text) is actually accessible to screen reader users.

PDF/UA is not itself a law. Legal obligations arise from regulations such as the German Accessibility Strengthening Act (BFSG) or BITV 2.0, which refer to technical requirements via the European standard EN 301 549. PDF/UA is the established way to neatly implement these requirements in PDF format.

PDF/UA-1 and PDF/UA-2: the two versions in comparison

The standard exists in two parts, which differ primarily in their technical basis:

feature PDF/UA-1 PDF/UA-2
standard ISO 14289-1 ISO 14289-2
Published 2012 2024
Technical basis PDF 1.7 (ISO 32000-1) PDF 2.0 (ISO 32000-2)
Tag model PDF 1.7 tag set with role mapping Modernized, expanded tag set with namespaces
Mathematical formulas No standardized award Structured markup possible via MathML

PDF/UA-1 (ISO 14289-1)

The first part of the standard appeared in 2012 and is based on PDF 1.7. PDF/UA-1 is still widely used today: most existing accessible PDFs and many checking tools are based on this version.

PDF/UA-2 (ISO 14289-2)

The second part was published in 2024 and is based on PDF 2.0. It uses its revised structural model: a more clearly defined tag set with namespaces, more precise rules for annotations and links, and the ability to mark mathematical formulas in a structured manner using MathML. For newly created documents, PDF/UA-2 is the current state of the art; PDF/UA-1 documents remain valid and usable.

What are the core requirements of PDF/UA?

Regardless of the version, PDF/UA essentially requires that all meaningful content be structured in a machine-readable manner. The most important requirements:

  • Tagged content (tag tree): Every component relevant to the content – ​​headings, paragraphs, lists, tables, figures, links – must be marked with the appropriate structural element. The tag tree represents the semantic structure of the document and is the basis for every assistive technology.
  • Logical reading order: The order of the tags must correspond to the reading sequence that makes sense in terms of content - even with multi-column layouts, marginal columns or information boxes. What is visually next to each other needs a clear sequence for the speech output.
  • Alt text for graphics: Images, diagrams and other graphical elements that convey information require alternative text that describes their content or function.
  • Marking of artifacts: Purely decorative elements - headers and footers, page numbers, background graphics, decorative lines - must be marked as artifacts so that screen readers can skip them instead of disrupting the flow of reading.
  • Document title and language: The document must have a meaningful title in the metadata, which is displayed instead of the file name. In addition, the main language must be declared - and language changes in the text must be marked so that the speech output reads out correctly.
  • Unicode mappability: Every character must be uniquely mapped to Unicode. This is the only way text-to-speech systems, search functions and copy functions can process the text reliably.
  • Correct table and list structures: Tables need excellent header cells and clear mappings between header and data cells; Lists need real list structures instead of just bulleted paragraphs.

How are PDF/UA and WCAG related?

In short: The WCAG define the success criteria, PDF/UA defines the technical implementation for the PDF format. The Web Content Accessibility Guidelines (WCAG) formulate technology-neutral requirements for digital content - this includes websites as well as documents. PDF/UA translates these requirements into concrete, verifiable specifications for PDF technology.

The two sets of rules do not replace each other, they complement each other: A PDF/UA-compliant document meets the structural requirements, while the WCAG introduces additional content criteria - such as sufficient color contrast or understandable link texts. In practice, the combination of PDF/UA compliance and WCAG checking is the reliable path to accessible PDF documents. Our article explains how the WCAG criteria should be specifically applied to PDFs WCAG for PDF documents.

What role do EN 301 549 and BFSG play?

The EN 301 549 is the European standard for the accessibility of information and communication technology. Chapter 10 deals with documents outside of websites - including PDFs - and adopts the relevant WCAG success criteria.

The BFSG implements the European Accessibility Act (EAA) into German law and has been applicable since June 28, 2025 for the products and services mentioned in Section 1 BFSG - for example in banking, e-commerce or passenger transport. So it doesn't affect all companies, but many: Where a service falls within the scope of application, the associated PDF documents - contracts, product information, invoices - must also be accessible. EN 301 549 serves as the relevant technical reference point. This also applies to public bodies BITV 2.0, which also refers to EN 301 549. Details on the scope and deadlines can be found in the article BFSG and accessible PDFs as well as on our page for Authorities and public administration.

How to check PDF/UA compliance?

A reliable test always combines two levels: machine validation and human evaluation.

Machine testing

Established tools like veraPDF or that PDF Accessibility Checker (PAC) validate the formal requirements automatically: Is a tag tree available? Are the document title and language set? Can all characters be mapped to Unicode? Are there alternative text entries missing? Such syntactic criteria can be reliably determined by machine.

The Matterhorn Protocol as a test catalog

The Matterhorn Protocol The PDF Association specifies the requirements of PDF/UA-1 in a structured test catalog with 31 test points and 136 individual error conditions. It explicitly distinguishes between conditions that software can check automatically and those that require human judgment. The protocol is therefore the common reference for testing tools and manual audits.

Human test

Whether an alternative text accurately describes the image content, whether the reading order makes sense in terms of content or whether a table structure reflects the actual meaning of the data - no purely mechanical test can decide this. People have to evaluate these semantic questions, ideally supplemented by a practical test with a screen reader.

How Accessful helps you with PDF/UA compliance

Accessful automatically audits and remediates PDF documents according to PDF/UA-2, WCAG 2.2 and EN 301 549. The AI ​​adds the structural level - tags, reading order, alt text, table and list structures - while the layout and content of your document remain unchanged. Results are typically available in seconds, even when batch processing thousands of documents.

For each remediated document, you will receive an audit-ready report as proof of audits and compliance requirements. The pages show how this works in detail PDF accessibility with AI and How it works – or you can test it directly with your own documents using the free PDF check.