Accessible PDFs

Accessible PDFs are tagged PDFs — documents with a structure tree describing the logical reading order and semantics of the content, independent of its visual layout. A screen reader uses that tree to read the document to a blind or low-vision user. Without it, the user gets either nothing or a jumbled mess.

EuroPDF generates tagged PDFs when you enable the PDF/UA profile as an API option. But the PDF is only as accessible as the HTML you send us. Prince’s own documentation is honest about this:

“Prince helps in creating accessible documents. Note, however, that Prince does not create WCAG-compliant documents by itself.” — PrinceXML docs

This page is about the HTML side: the semantic patterns that let Prince produce a tagged PDF that passes PDF/UA checkers. You write proper HTML, we handle the rest.

A complete example

▶ Open in Playground

<!DOCTYPE html>
<html lang="en">
<head>
  <title>Quarterly Report 2026</title>
</head>
<body>
  <header>
    <h1>Quarterly Report 2026</h1>
  </header>
  <main>
    <h2>Summary</h2>
    <p>Revenue increased 18% year-over-year...</p>

    <h2>Revenue by Region</h2>
    <table>
      <caption>Revenue in EUR, Q1 2026</caption>
      <thead>
        <tr><th scope="col">Region</th><th scope="col">Revenue</th></tr>
      </thead>
      <tbody>
        <tr><th scope="row">Germany</th><td>€480,000</td></tr>
      </tbody>
    </table>
  </main>
  <footer>
    <p>Confidential. Internal use only.</p>
  </footer>
</body>
</html>

Everything that matters for accessibility is in that markup: a lang attribute on <html>, a document <title>, semantic landmarks (<header>, <main>, <footer>), a proper heading hierarchy, and a table with <caption>, <th scope>, and explicit <thead>/<tbody>. No special classes, no ARIA attributes, no CSS tricks — just well-structured HTML.

What EuroPDF does for you

Tags the PDF automatically when you output with the PDF/UA profile, mapping your HTML elements to the correct PDF structure tags.
Preserves the reading order from the HTML DOM, which is what assistive tech follows.
Embeds the document language in the PDF metadata for screen readers.
Carries alt text from <img alt="..."> into the PDF image structure.
Emits a tagged structure tree that conforms to the PDF/UA (ISO 14289) and PDF/A-3a specifications.

Try it now

Test it instantly in the Playground — no sign-up required.

Open in Playground Sign up for free

More examples

Language declaration

▶ Open in Playground

<html lang="de">
  <body>
    <p>Dieses Dokument ist auf Deutsch.</p>
    <p lang="en">This paragraph is in English.</p>
  </body>
</html>

Without lang, screen readers guess the language (usually wrong) and mispronounce the text. Mark the document language once on <html>, and override it with lang="..." on any element whose content is in a different language.

Images with alt text

▶ Open in Playground

<figure>
  <img src="chart.png" alt="Bar chart showing Q1 2026 revenue of 1.2M, up 18% YoY">
  <figcaption>Figure 1: Quarterly revenue</figcaption>
</figure>

alt describes what the image conveys, not what it looks like. For decorative images (spacers, dividers), use alt="" — not missing alt, which reads the filename.

Accessible tables

▶ Open in Playground

<table>
  <caption>Team headcount by department</caption>
  <thead>
    <tr><th scope="col">Department</th><th scope="col">Headcount</th></tr>
  </thead>
  <tbody>
    <tr><th scope="row">Engineering</th><td>42</td></tr>
    <tr><th scope="row">Product</th><td>12</td></tr>
  </tbody>
</table>

The scope attributes let assistive tech announce the right row and column headers as the user navigates between cells.

Correct heading hierarchy

▶ Open in Playground

<h1>Document title</h1>
<h2>Chapter One</h2>
<h3>A subsection</h3>
<h2>Chapter Two</h2>

One <h1> per document, no skipped levels. Use CSS to adjust visual size — never skip levels to get a smaller look.

Semantic landmarks

▶ Open in Playground

<header><h1>Employee Handbook</h1></header>
<nav>Contents: Introduction · Policies · Benefits</nav>
<main>
  <h2>Introduction</h2>
  <p>...</p>
</main>
<aside>Tip: Use only one main element per document.</aside>
<footer>© 2026 Internal document.</footer>

Landmark elements let screen reader users jump directly to the content area and skip repetitive navigation.

Semantic lists

▶ Open in Playground

<ul>
  <li>Unordered bullet list — order doesn't matter</li>
</ul>

<ol>
  <li>Ordered list — sequence is meaningful</li>
  <li>Steps, rankings, timelines</li>
</ol>

<dl>
  <dt>Tagged PDF</dt>
  <dd>A PDF with a structure tree for assistive tech.</dd>
</dl>

Never fake a list with paragraphs and bullet characters — assistive tech loses the list structure entirely and reads it as flowing prose.

Form labels

▶ Open in Playground

<label for="email">Email address</label>
<input type="email" id="email" name="email">

<label>
  Phone
  <input type="tel" name="phone">
</label>

Every form field needs an associated label — either via for/id or by nesting the input inside the <label>. Without this association, assistive tech has no way to announce what each field is for.

EuroPDF wraps PrinceXML — the full reference for PDF/UA output, tagged PDF, and PDF profiles lives in the Prince user guide.