Accessible PDFs: start with the source document

What makes a PDF accessible?

A PDF can look perfect to the eye and still be a jumbled stream of text for a screen reader. What matters are the tags: an invisible structure that says what is a heading, paragraph, list, table, image or link, and in which order the content is read. Many PDF programs show in the document properties whether a PDF is tagged. Selectable text alone, then, does not mean a PDF is accessible.

In its PDF techniques for WCAG, the W3C describes how such requirements are implemented: text alternatives for images, heading tags, correct reading order, document title, language setting, table and list tags, bookmarks and labelled form fields.

Does a PDF on my website have to be accessible?

That depends on the legal framework. Public bodies follow BITV 2.0 (the German federal regulation on accessible information technology) or the state rules. In § 1(4), the BFSG excludes office file formats published before 28 June 2025. This exemption does not apply to newer documents. For older documents that you continue to revise, you should not rely on it without checking.

If in doubt, have an expert check whether a specific document is subject to an obligation; this article is not legal advice. Regardless of that, a PDF with important information should be readable by everyone.

How do you prepare the source document?

Office programs such as Word come with a built-in accessibility checker. It finds gaps such as missing text alternatives early, but it also only checks part of the picture. What you set up correctly in the source document, you do not have to laboriously fix in the PDF after export.

1. 1Use real headings and lists via styles instead of visual formatting alone, such as bold and larger text.
2. 2Give the document a meaningful title and the correct language.
3. 3Add image descriptions and clear link text.
4. 4Design tables with clear headers and a simple structure, without merged cells where you can avoid them.
5. 5Check contrast and avoid pure image scans without accessible text.
6. 6Set the order of text boxes, footnotes and sidebars deliberately, especially in layout programs such as InDesign.
7. 7When exporting, make sure tags and the document structure are included. Printing to a PDF file often loses this information.

How do you check a PDF after export?

Check the generated tags and the logical reading order, especially with multi-column layouts, text boxes and footnotes. Check that the window title shows the document title rather than the file name and that the language is set.

Check links and forms. In tables, header cells must be tagged as such so that a screen reader names the right column heading when reading a cell. Decorative elements such as header lines should be marked as artefacts so that they are not read out. Long documents need bookmarks. Combine suitable PDF checking tools with a manual review and a test with assistive technologies. In the screen reader test, you listen to the beginning, jump through the headings and read a table cell by cell.

What do you do with scanned PDFs?

A scanned document is initially just an image. A screen reader finds no text in it. Optical character recognition (OCR) produces real text and is a first step; the W3C describes it as a technique of its own.

But recognition makes mistakes, especially with tables, handwriting and poor originals. Check the result and add structure tags. A screen reader reads out every recognition error. Mark signatures, stamps or logos as artefacts, or give them a short text alternative if they carry information. For important documents, it is often quicker and more reliable to recreate the source file.

How do you make PDF forms accessible?

Form fields in a PDF need a label that assistive technologies can read, as well as a name, role and value. Required fields must be marked, and incorrect input needs a clear message.

This takes effort and works more or less well depending on the PDF program. If people need to send you data, an accessible HTML form on your website is often the better solution. If you still offer a PDF form, test filling it in with the keyboard and a screen reader.

When is an HTML page the better choice?

If content works well as an HTML page, offer it directly on your website. HTML adapts to screen size and zoom and is easier to keep up to date. For downloads that are needed, the linking page should clearly describe the file type and content, for example ‘Price list 2026 (PDF)’. An HTML alternative can create additional ways of access. For long reports, even an HTML summary that links to the full PDF helps.

A PDF makes most sense when a document is meant to be printed, archived or passed on unchanged, such as a contract or a form to be signed.

What are typical PDF mistakes?

1. 1Untagged PDF: The content has no structure and is read out in random order.
2. 2The window title shows the file name instead of a document title.
3. 3The language is missing or wrong, so a German text is read out with English pronunciation.
4. 4Headings are only formatted in bold and not tagged as headings.
5. 5Tables are images or have no header cells.
6. 6Long documents without bookmarks, where you cannot jump to chapters.
7. 7Page numbers and headers are read out in the middle of the text on every page.
8. 8Links appear only as a long URL or as ‘here’ with no recognisable destination.
9. 9Important notices are highlighted only by colour.
10. 10Form fields without labels, which the screen reader only announces as ‘edit field’.