Definition: PDF/UA ("Universal Accessibility") is the international standard ISO 14289-1 for accessible PDF documents [4]. It sets out how a PDF must be built so that assistive technology can read it: tags for the structure, a logical reading order, alt text and the document's language. PDF/UA-2 (ISO 14289-2) extends it to PDF 2.0.
German: PDF/UA · Also: accessible PDF, tagged PDF · Last reviewed: 27 September 2026
What it is about
On screen, a PDF looks like printed paper. A screen reader does not care how it looks; it needs the structure underneath. What is a heading, what is a list, what is a table, and in what order should it be read? In a PDF, that structure lives in tags. To many people, an untagged PDF is a page of characters with no order.
PDF/UA describes what those tags and the rest of the information must look like. It addresses both sides: the software that creates PDFs and the software that reads them.
What PDF/UA requires
- All content is tagged: headings, paragraphs, lists, tables, figures, links. Pure decoration is marked as an artifact so it is not read out.
- Reading order follows the tag structure, not the position on the page.
- Figures have alternative text, like alt text in HTML.
- Tables have header cells, so data cells can be related to them.
- The document language is set.
- A document title is set and shown instead of the file name.
- Text is real text with embedded fonts, not an image of a scanned page.
PDF/UA and WCAG
WCAG applies to web content, and a PDF on a website is part of that. The W3C publishes techniques specifically for PDF, for example on alt text and headings [1][2]. EN 301 549, the European standard, sets out requirements for documents that are not web pages in chapter 10, largely following WCAG [3]. PDF/UA, by contrast, describes the technology of the file. The two complement each other: a PDF can conform to PDF/UA and still carry useless alt text.
How to check a PDF
Free checkers such as PAC or veraPDF test the part of PDF/UA a machine can decide: missing tags, a missing language, fonts that are not embedded. The PDF Association's Matterhorn Protocol lists the checkpoints and separates what software can test from what needs a person. Whether alt text fits, whether the reading order makes sense and whether headings have the right level is a human judgement – best made with a screen reader.
Common mistakes
- Scanned documents with no text recognition: to a screen reader the page is empty.
- Tagged automatically, never checked: the export adds tags, in the wrong order.
- Word documents without styles: bold lines never become headings.
- Tables as images.
- A PDF as the only route to important information, when an HTML page would be easier to make accessible.
How Reviseberg handles it
Reviseberg checks only a small part of PDF/UA: the crawl opens the PDFs your pages link to on your own site and checks whether they are tagged, whether the tags hold content, the document language, a displayed title and figure alternative text. It does not check whether the reading order is right, tables, contrast, fonts or form fields. For a full PDF/UA check you need a dedicated checker and a manual review. Reviseberg says so plainly rather than presenting a site full of PDFs as tested.
Related terms
Alt text · Heading structure · Screen reader · Language attribute (lang) · EN 301 549 · Manual accessibility testing
Further reading
- WCAG 1.1.1 Non-text Content
- WCAG 1.3.1 Info and Relationships
- EN 301 549 explained