Accessible PDFs — what does the law require?
Accessibility requirements don't stop at the web pages — they cover the documents published on them. A PDF that looks perfect on screen can be completely mute to a screen reader: without tags there is no structure to navigate by, and without a text layer there is nothing to read aloud at all. Here we walk through what makes a PDF accessible, what the law requires and the most common issues we find.
What does the law say?
For the Swedish public sector, the Act on Accessibility to Digital Public Service (DOS-lagen) applies: documents published after 23 September 2018 must be accessible. Older documents are also covered if they are needed for active administrative processes — an old form still in use doesn't get a pass because of its age.
For private businesses covered by the EU's European Accessibility Act, the Swedish Accessibility Act (2023:254) applies since 28 June 2025 — e-commerce, banking services, passenger transport and electronic communications among others, with services to consumers at the core. The required level is the same in both cases: the European standard EN 301 549, which applies WCAG at levels A and AA — and it covers documents, not just web pages.
What makes a PDF accessible?
- Tagged structure — headings, lists, tables and reading order are marked up in the document, so a screen reader can navigate it like a web page instead of facing an unformatted wall of text.
- Document language — the language is set in the document's properties, so the screen reader picks the right pronunciation. A Swedish text read aloud with English pronunciation rules is next to incomprehensible.
- Title — the document has a descriptive title in its metadata, not just a file name. The title is the first thing the user hears.
- Text layer — the text exists as real text, not just as a scanned image. An image scan without a text layer can't be read aloud, searched or selected.
Common issues we see
- Scanned documents without a text layer — minutes, forms and contracts scanned to image. To a screen reader, the document is empty.
- PDFs exported without tags — the tagged-PDF checkbox was never ticked at export, so the structure from the source document was lost along the way.
- Missing or meaningless title — the document introduces itself as "finalversion-FINAL2.pdf".
- Document language missing or left at the export tool's default.
- Old documents still used in active matters but never remediated.
How CompliantHQ tests this
The base scan — included in the trial as well — finds PDFs linked from the scanned pages and reviews each document: is it tagged, is the document language set, is there a title, and is the content real text or just a scanned image.
The result shows which documents fail and on what — so you can prioritise: fix the documents that are actually used first, and remove the ones no longer needed.
How to fix it
- Fix the source, not just the PDF — set headings, lists and tables with the formatting tools in Word or InDesign, and choose tagged PDF at export. Then the next version comes out right automatically.
- Set the document title and language in the document properties before exporting.
- Redo scanned documents: start from the original file if it still exists, otherwise OCR the scan and check the result.
- Ask whether the content needs to be a PDF at all — a regular web page is often easier to make accessible and easier to keep up to date.
- Clean up: old documents that are no longer needed can be unpublished instead of remediated.
What the check covers
- That PDF documents are accessible: tagging, document language, title, and whether the document is just a scanned image.
WCAG criteria covered
- 2.4.2 Page titled — every page must have a title describing what it's about. It's the first thing a screen reader announces, and what shows in the browser tab. (A)
- 3.1.1 Language of page — the page's language must be declared in the code. Otherwise a screen reader may read Swedish text with English pronunciation — incomprehensible to the listener. (A)
Common questions
Do we have to fix old PDFs?
In the public sector, DOS-lagen covers documents published after 23 September 2018. Older documents are covered if they are needed for active administrative processes — if the form is still in use, it must be accessible, regardless of age.
What is a tagged PDF?
A PDF with built-in structure: headings, lists, tables and reading order are marked up as tags in the document. The tags aren't visible in the layout, but they are what the screen reader navigates by.
Our PDF lets you search and select text — is that enough?
No, the text layer is just the foundation. Without tags the document lacks structure, and without a document language and title it still fails. But it's a good start — a scanned image without a text layer is a bigger problem.
Is it better to publish the content as a web page instead?
Often, yes. A web page is easier to make accessible, works better on mobile and can be updated without a new export. PDF makes sense when the document needs to be printed or archived in exact form.
Want to see what we find on your site?
Run a free scan — all four modules included for 30 days, no card required.