What WCAG 2.1 AA means for documents, not just web pages
WCAG 2.1 is written for all web content, and US regulators have made Level AA the enforceable standard for public-sector documents: the ADA Title II rule and the HHS Section 504 rule both apply it to PDFs, word processor files, presentations and spreadsheets. This page maps the success criteria to what they actually require inside a document.
The standard
Does WCAG apply to PDFs and other documents?
Yes. WCAG is technology-neutral: its success criteria apply to any content delivered over the web, and a document opened from a website is evaluated the way a web page is. The DOJ's ADA Title II rule and the HHS Section 504 rule both expressly extend WCAG 2.1 Level AA to conventional electronic documents.
W3C publishes PDF-specific techniques(opens in new tab) for meeting the criteria, and the regulatory definitions of web content name documents explicitly — PDF, word processor, presentation and spreadsheet files. So 'it's a PDF, not a page' changes the techniques, never the obligation.
Success criteria
Which WCAG 2.1 AA success criteria apply to a PDF?
Eight criteria account for most document failures. Each maps to a checkable property of the file — tags, reading order, metadata, text layer — and the last column says how much a checker can settle.
| Criterion | What it requires in a document | Machine-checkable? |
|---|---|---|
| 1.1.1 Non-text content — alternative text(opens in new tab) | Every meaningful image, chart and figure needs alternative text conveying its purpose; purely decorative graphics are marked as artifacts so assistive technology skips them. | Partly — presence yes, meaning no |
| 1.3.1 Info and relationships — tag structure(opens in new tab) | Headings, lists, tables and paragraphs must be expressed in the document's tag tree, not just visually: real heading levels, list structures, and table cells associated with their header cells. | Largely — structure yes; the right structure takes a reader |
| 1.3.2 Meaningful sequence — reading order(opens in new tab) | The order assistive technology reads the content must match the logical order a sighted reader follows — multi-column layouts, sidebars and callouts are where untagged PDFs fail. | Partly — an order yes; sense no |
| 1.4.3 Contrast (minimum)(opens in new tab) | Text needs a contrast ratio of at least 4.5:1 against its background (3:1 for large text) — a design-time property that remediation can flag but authors must fix. | Partly — measurable on live text; outside PDF/UA-1, most checkers skip it |
| 1.4.5 Images of text — scanned pages(opens in new tab) | A scanned page is one big image of text and fails outright: it needs OCR to restore a real text layer, then tagging, before any other criterion can be met. | Yes — an image-only page is detectable |
| 2.4.2 Page titled — document title(opens in new tab) | The file needs a descriptive title in its metadata, configured to display instead of the file name — 'scan_final_v3.pdf' in a screen reader's window list helps no one. | Yes — present and set to display |
| 3.1.1 Language of page(opens in new tab)3.1.2 Language of parts(opens in new tab) | The document's language must be declared so screen readers pick the right voice and pronunciation; passages in another language are marked at the tag level (3.1.2). | Largely — the document language yes; unmarked passages take a reader |
| 4.1.2 Name, role, value — form fields(opens in new tab)3.3.2 Labels or instructions(opens in new tab) | Every form field needs a programmatic label, a correct role and a set value; visible instructions and required-field indicators must also be available to assistive technology (3.3.2). | Partly — presence yes; whether the label matches the question, no |
The right-hand column is the whole argument for verification: two criteria a checker settles outright, two largely, four only half — and those four are where documents that 'pass' are still unusable.
Conformance
How is WCAG conformance assessed for non-HTML content?
Per document, the way a web page is assessed: every applicable Level A and AA success criterion must be satisfied for the file to conform. Automated checkers verify the machine-testable subset — tags present, title and language set, a text layer that exists — while reading-order sense and alt-text quality require human judgment.
WCAG's conformance model evaluates full pages and complete processes, so a form that conforms except for its final signature page does not conform. For PDF specifically, the PDF/UA standard defines the file-format requirements machine checkers can verify reliably — the practical companion to WCAG's outcome requirements.
Legal status
Which laws require WCAG 2.1 AA for documents?
WCAG 2.1 is a W3C Recommendation, not a law. What makes it enforceable is the regulation that adopts it — and each regime below has its own regulator: the DOJ, HHS, the US Access Board, Ontario, or an EU member state.
| Regime | WCAG version | Covers | When |
|---|---|---|---|
| ADA Title II | 2.1 AA | US state and local governments | April 26, 2027 (50,000 or more people) · April 26, 2028 |
| HHS Section 504 | 2.1 AA | Recipients of HHS federal financial assistance | May 11, 2027 (15 or more employees) · May 10, 2028 |
| Section 508 | 2.0 AA | US federal agencies and their ICT vendors | Since January 18, 2018 |
| AODA (Ontario) | 2.0 AA | Ontario public sector; organizations with 50 or more employees | Since January 1, 2021 |
| European Accessibility Act | 2.1 AAvia EN 301 549 | Products and services in the EAA's scope | Since June 28, 2025 |
Because WCAG versions are backwards-compatible, remediating documents to 2.1 AA satisfies every 2.0-based regime at the same time. Each guide states its own cohorts and exceptions.
The plan
How do you make a document library conform?
Start from an inventory, not a file-by-file guess: enumerate the documents on your domains, classify them, and let the currently-used ones jump the queue. Remediate to WCAG 2.1 AA and PDF/UA together — the same tagging work satisfies both — and demand a verification report per file so conformance is evidence, not assertion.
Inventory: enumerate every PDF and Office document on your public domains.
Classify: separate scanned files, forms and complex layouts — they need different remediation levels.
Remediate: fix tags, reading order, alt text, titles, language and form labels to 2.1 AA.
Verify: run machine checks plus human review, and keep the report with the file.
FAQ
WCAG for documents, answered
Buying for a larger organization?
Book a 20-minute demo(opens in new tab)Does WCAG apply to Word, PowerPoint and Excel files, or only PDFs?
It applies to all of them. The US rules define covered web content to include word processor, presentation and spreadsheet files alongside PDFs. The techniques differ — Office formats carry structure through styles and built-in headings — but the success criteria and the Level AA bar are identical.
Do documents have to meet Level AAA?
No. The ADA Title II and HHS Section 504 rules require Level AA, which includes all Level A criteria. Level AAA is optional everywhere: some AAA criteria, like the enhanced 7:1 contrast ratio, are worthwhile targets for key documents, but no US rule mandates them.
Can a scanned PDF conform to WCAG?
Not as an image. A scan fails non-text content (1.1.1) and images of text (1.4.5) immediately, because there is no real text for assistive technology to read. After OCR restores the text layer and the file is fully tagged with a correct reading order, it can conform like any other document.
Can an automated checker prove a document conforms?
No — machine checks are the floor, not the ceiling. Automated tools verify that tags, a title and a language exist and that measurable ratios pass, but they cannot judge whether the reading order makes sense or the alt text says the right thing. Conformance takes machine checks plus human review.
What about WCAG 2.2 — should documents target it instead?
The US rules cite WCAG 2.1 Level AA specifically, so that is the compliance bar. WCAG 2.2 builds on 2.1 and most of its additions concern interactive interfaces rather than documents; targeting 2.1 AA plus PDF/UA remains the practical standard for document remediation today.
Sources
Every regulatory claim on this page is cited to one of these primary sources.
- W3C — Web Content Accessibility Guidelines (WCAG) 2.1
- W3C — Techniques for WCAG 2.1: PDF techniques
- W3C — Understanding WCAG 2.1: conformance requirements
- Federal Register, April 24, 2024 — ADA Title II web accessibility final rule (89 FR 31320), adopting WCAG 2.1 AA
- Federal Register, April 20, 2026 — Interim final rule extending the ADA Title II compliance dates (91 FR 20902)
- Federal Register, May 9, 2024 — HHS Section 504 final rule (89 FR 40066), adopting WCAG 2.1 AA
- Federal Register, May 11, 2026 — Interim final rule extending the HHS Section 504 compliance dates (91 FR 25496)
- US Access Board — Revised Section 508 Standards (Information and Communication Technology)
- Ontario.ca — How to make websites accessible (AODA)
- EUR-Lex — Directive (EU) 2019/882, the European Accessibility Act
Changelog
- Criteria and regime tables added with catalog dates; W3C techniques, conformance requirements and the HHS rule cited; verification, human review and the site scan linked.
- Initial publication.
Reviewed quarterly, and whenever a cited source changes. Next review: December 18, 2026.