ADA Title II and PDFs: the 2027 deadline
The Department of Justice’s Title II web rule is the reason accessibility suddenly has a date attached. It covers electronic documents, not just web pages — which means every PDF a public entity publishes is in scope.
Last reviewed 25 July 2026
What the rule requires
In 2024 the DOJ finalised a rule under Title II of the Americans with Disabilities Act requiring state and local government web content and mobile apps to conform to WCAG 2.1 Level AA. The scope is broader than most people assume: it names web pages, mobile applications, electronic documents and multimedia.
That last category is where most organisations are exposed. A council can rebuild its website and still fail, because the site links to thousands of untagged PDFs — agendas, minutes, budgets, permit forms, school handbooks — and each of those is web content under the rule.
The dates
The original compliance dates were April 2026 and April 2027. On 20 April 2026 the DOJ issued an interim final rule extending both by a year:
| Who | Deadline |
|---|---|
| Public entities serving 50,000 or more people | 26 April 2027 |
| Public entities serving fewer than 50,000, and special district governments | 26 April 2028 |
An extension is not a reprieve. The rule took effect in June 2024; only the compliance dates moved. Title II obligations and private litigation risk did not pause in the meantime, and separately, the HHS Section 504 rule carried a May 2026 date of its own for recipients of federal health funding.
What makes a PDF conform
A PDF that looks fine on screen can be completely opaque to a screen reader. Conformance is about structure underneath the visual layer, and it comes down to a handful of things:
- Tagged content. The file has to declare itself tagged, and carry a structure tree that marks headings, paragraphs, lists and tables as what they are.
- Reading order. The order the tags run in has to match the order a human would read. A two-column layout tagged straight across produces fluent nonsense.
- Alt text. Images that carry meaning need descriptions; purely decorative ones must be marked as artifacts so they are skipped rather than announced.
- Document language and title. The language drives screen-reader pronunciation; the title is what gets announced instead of a filename.
- No hostile permissions. Some PDFs set flags that block text extraction outright, which blocks assistive technology along with it.
The PDF-specific standard for all of this is PDF/UA-1 (ISO 14289-1). WCAG is written for web content generally; PDF/UA is how those requirements land in a PDF, and it is what validators actually test against.
How to find out where you stand
Check a representative sample before you estimate anything. Pull ten documents that matter — the ones people actually download — and test them. Most organisations find that nearly all of them fail, which is useful information precisely because it is unambiguous.
Automated checking catches the machine-testable layer: tagging, language, title, alt-text presence, permissions. It cannot judge whether an alt text is well written, or whether a reading order makes sense to a person. Both matter, and a tool that claims to settle the second kind automatically is overselling.
Check one of yours
Free, runs entirely in your browser, and the file is never uploaded — which matters if the document isn’t public yet. It reports the seven machine-checkable PDF/UA requirements and tells you which ones fail.
Check a PDF freeThe European Accessibility Act
If you sell into the EU, a second clock has already run out. The European Accessibility Act applied from 28 June 2025, and it reaches private companies rather than only public bodies. Digital content is measured against EN 301 549, which is itself built on WCAG 2.1 Level AA — so the technical target is the same one Title II names.
Enforcement has been ramping through 2026, with the first lawsuits filed in France in November 2025 and other member states following. Penalties are set nationally and vary widely, from roughly €60,000 in Ireland to around €900,000 in Sweden. The Act binds non-EU companies selling into the bloc, and exempts only micro-enterprises — fewer than 10 staff and under €2m turnover.
Remediation, realistically
Doing this by hand in Acrobat Pro is entirely possible and genuinely slow — tagging a complex document is measured in hours, not minutes. Agencies charge per page for the same reason, typically in the region of $5–8, with rush work costing substantially more. Enterprise remediation software runs to four figures a seat per year on top of an Acrobat licence.
Automation handles the structural layer well: tagging, reading order, language, metadata, and the PDF/UA identifier. It handles judgement poorly. A sensible plan automates the bulk, then spends human time on the documents where meaning is carried by images and complex tables.
Whatever route you take, insist on validator output rather than an assurance. veraPDF is the open-source reference validator for PDF/UA, and a report from it is the closest thing to objective evidence that a file conforms.
This is a plain-language summary for people planning work, not legal advice. Deadlines, thresholds and enforcement differ by jurisdiction and by entity type — confirm your own obligations with counsel.