How to make PDFs WCAG 2.1 AA compliant
The actual steps to remediate a PDF for ADA Title II, what automated tools do and don't handle, and how to validate the result.
What "accessible" means for a PDF
For ADA Title II, the target is WCAG 2.1 Level AA applied to the document itself (Source: 28 CFR ยง 35.200 ) . In PDF terms that maps closely to PDF/UA, the technical standard for accessible PDFs. A compliant PDF isn't just one that looks right on screen; it has an underlying structure that assistive technology can read, navigate, and understand.
Start from the source when you can
The cleanest path is usually to fix accessibility in the source file (Word, InDesign, Google Docs) and then export to a tagged PDF, because the authoring tool can carry structure into the export. For documents you no longer have a source for, or scanned PDFs, you remediate the PDF directly, which takes longer.
The remediation steps
These are the elements that determine whether a PDF actually works with assistive technology:
- Tag structure. Give the document a real tag tree: headings (H1-H6), paragraphs, lists, and so on, so a screen reader can move through it by structure.
- Reading order. Make sure the tagged order matches the logical reading order, not just the visual position on the page. Multi-column layouts and text boxes are common traps.
- Alternative text. Add concise alt text to meaningful images, charts, and logos, and mark purely decorative images as artifacts so they're skipped.
- Tables. Tag data tables with header rows and cells so relationships are announced. Avoid using tables purely for visual layout.
- Links. Use descriptive link text rather than bare URLs, and make sure links are tagged so they're reachable by keyboard.
- Language and title. Set the document language and a meaningful title in the file properties, and set the title to display in the window.
- Color and contrast. Meet contrast minimums for text, and never rely on color alone to convey meaning.
- Forms. For interactive PDFs, label every field, set a logical tab order, and add instructions where needed.
Automated tagging tools (including the auto-tag feature in PDF editors) can do a first pass on some of these, but they generally leave reading-order, alt-text, and table issues that need a human to finish. That's the gap that tends to catch teams off guard.
How to validate the result
A document isn't done because a checker shows green. The reliable approach combines:
- An automated check against WCAG 2.1 AA and PDF/UA to catch the mechanical failures.
- A manual pass with a screen reader and keyboard, which is the only way to confirm reading order, link clarity, and that the document actually makes sense non-visually.
Keeping a short, dated record of how each document was tested generally helps both your compliance file and any future review.
Doing it at scale
One document is a manageable afternoon. A library of hundreds or thousands is a different problem, and it's where most public entities get stuck. At scale the work usually needs consistent tooling, a triage order (see which to prioritize), authoring templates so new documents go out accessible, and human validation so the output actually holds up.
If you want a quick way to gauge where your documents stand, the readiness checklist walks through the same points as a yes/no check.