Skip to content
Layout Guide
Accessible PDF

Accessible PDF documents (PDF/UA)

PDF/UA (ISO 14289, “Universal Accessibility”) is the standard for accessible, tagged PDF. A PDF/UA document carries a logical structure tree (headings, paragraphs, lists, figures, formulas, …) that assistive technology such as screen readers uses to convey the document in a meaningful reading order.

XTS can produce PDF/UA-1 and PDF/UA-2 output. This chapter explains how to turn it on, what your layout has to provide, and how to verify the result.

Turning on conformance

The output conformance is configured on the command line or in the configuration file.

# PDF/UA-2 (based on PDF 2.0)
xts --pdfua 2
pdfua = "2"

There are three orthogonal conformance axes; each is set independently:

Axis Values Standard
pdfua none, 1, 2 PDF/UA (accessibility)
pdfa none, 3b PDF/A (archiving)
pdfx none, X-3, X-4 PDF/X (printing)

Axes from different families can be combined as long as they agree on the base PDF version, for example an accessible and archivable PDF:

xts --pdfa 3b --pdfua 1
Version coupling. PDF/UA-2 requires PDF 2.0, while PDF/A-3 and PDF/X-3/4 are PDF 1.x. XTS rejects contradictory combinations such as --pdfua 2 --pdfa 3b with a clear error. Pair PDF/UA-1 with PDF/A-3 (both PDF 1.x), or use PDF/UA-2 on its own.

What an accessible document needs

Turning on pdfua makes XTS emit the structure tree, marked content, language and metadata automatically. A few things, however, have to come from your layout and data.

A document title

PDF/UA requires a title that is shown instead of the file name. Set it with <PDFOptions>:

<PDFOptions title="Annual Report 2026"/>
If no title is set by the time the metadata is written, XTS falls back to a placeholder title and prints a warning. The document stays conformant, but you should always set a real, meaningful title.

A document language

Declare the natural language of the document with <Options>:

<Options mainlanguage="en"/>

Logical structure: headings, lists, figures, formulas

The structure tree is built from the HTML-like content you place with <HTML>. Use real semantic elements — headings <h1>–<h6>, paragraphs <p>, lists <ul>/<ol>, tables — and they become the corresponding structure elements.

<HTML>
  <h1>Quarterly figures</h1>
  <p>Revenue rose in all regions.</p>
  <ul>
    <li>Europe: +4%</li>
    <li>Asia: +7%</li>
  </ul>
</HTML>

Images must carry alternative text so they are tagged as a Figure with /Alt:

<HTML>
  <img src="chart.pdf" alt="Bar chart of revenue per region"/>
</HTML>

Formulas are written as MathML inside <HTML>. They are tagged as a Formula structure element with an alternative text and the original MathML attached as an associated file. Inline and block (display="block") formulas are both supported. A MathML formula needs a font with an OpenType MATH table, declared in CSS:

<StyleSheet>
  @font-face { font-family: "Latin Modern Math"; src: url("latinmodern-math.otf"); }
  math { font-family: "Latin Modern Math"; }
</StyleSheet>
<HTML>
  <p>The Pythagorean theorem is
    <math><msup><mi>a</mi><mn>2</mn></msup><mo>+</mo><msup><mi>b</mi><mn>2</mn></msup><mo>=</mo><msup><mi>c</mi><mn>2</mn></msup></math>.</p>
</HTML>

A complete minimal example 

pdfua = "2"
<Layout xmlns="urn:speedata.de/2021/xts/en">
  <Options mainlanguage="en"/>
  <PDFOptions title="Accessible report"/>

  <Record match="data">
    <PlaceObject>
      <TextBlock>
        <HTML>
          <h1>Accessible report</h1>
          <p>This document is tagged for accessibility.</p>
        </HTML>
      </TextBlock>
    </PlaceObject>
  </Record>
</Layout>
xts

Verifying accessibility

Generating a PDF/UA document does not guarantee it is meaningfully accessible — alt texts have to be sensible, the reading order has to make sense, and so on. Always verify the result.

The easiest way is the online checker at pdfuacheck.speedata.de: upload your PDF and it reports, per checkpoint, what passes and what fails, including the document’s structure tree.

Other established tools:

  • veraPDF — the reference open-source PDF/UA and PDF/A validator.
  • PAC (PDF Accessibility Checker) — a free Windows tool from the PDF/UA Foundation with a visual structure inspector.

A typical failure and its cause:

Report Cause
No dc:title / title missing No title on <PDFOptions>
Document language not set No mainlanguage on <Options>
Figure without alternative text <img> without an alt attribute
Untagged content Content placed outside the tagged <HTML> flow

See also