Skip to main content
WCAG · PDF accessibility

WCAG for PDFs: one standard, two formats

WCAG is not just for web pages - laws apply it to documents too. How the criteria translate into PDF tags, reading order and forms, and where PDF/UA completes the picture.

ADA compliance guide
Yes, WCAG applies to PDF files. The W3C defines WCAG for all "web content", and the laws that reference it - EN 301 549 (its Chapter 10 applies WCAG criteria to non-web documents), Section 508, the ADA - cover the documents you publish, not only your HTML pages. Alongside WCAG sits PDF/UA (ISO 14289), the PDF-specific technical standard: WCAG says what an accessible outcome is, PDF/UA says how a PDF file must be built internally to deliver it. An accessible PDF program targets both: WCAG 2.1/2.2 Level AA for legal conformance, PDF/UA for technical completeness - and the foundation of both is a correctly tagged PDF.

Does WCAG apply to PDF files?

It does, on two levels. First, the W3C wrote WCAG for "content" generally and publishes a dedicated set of PDF Techniques showing how to satisfy the criteria in PDF documents. Second - and more importantly for compliance - the legal instruments that reference WCAG apply it to documents:

The practical consequence: if your compliance program tests HTML pages and skips the document library, the most-downloaded files on the site - invoices, contracts, guides - are usually its biggest open risk.

WCAG vs PDF/UA: what each one covers

PDF/UA (ISO 14289) is the PDF-native accessibility standard, published in two parts: PDF/UA-1 (ISO 14289-1, for PDF 1.7) and PDF/UA-2 (ISO 14289-2, for PDF 2.0). It is not a competitor to WCAG - it is the file-format layer underneath it:

WCAG 2.1 / 2.2PDF/UA-1 and PDF/UA-2
What it definesOutcome requirements for any content: perceivable, operable, understandable, robustHow a PDF must be built internally: tag structure, reading order, Unicode mapping, metadata, embedded fonts
Covers contrast, media, reflowYes - 1.4.3, 1.4.10 and the restNo - PDF/UA leaves visual presentation to WCAG
Covers tag-tree correctnessIndirectly, via 1.3.1 and 1.3.2Yes - in exhaustive, machine-verifiable detail
Referenced by lawEN 301 549, EAA, ADA rule, Section 508Referenced as the technical method (EN 301 549 cites it as a way to satisfy Chapter 10)
ConformancePer level: A / AA / AAABinary: a file conforms or it does not

The working rule: target WCAG Level AA as the legal bar, and validate against PDF/UA (both UA-1 and UA-2) as the engineering bar. A file can pass PDF/UA and still fail WCAG on contrast; it can look WCAG-friendly and be unreadable to a screen reader because the tags are wrong. Testing both closes both gaps.

How WCAG criteria map to a PDF

The same criteria you test on a web page take a specific technical form inside a PDF. The high-impact mappings:

WCAG criterionIn a PDF file, this means
1.1.1 Non-text ContentAlternate text on every Figure tag; decorative graphics marked as Artifacts
1.3.1 Info and RelationshipsA complete tag tree: H1-H6, P, lists as L/LI, tables with TH/TD and headers
1.3.2 Meaningful SequenceTag order = reading order, especially in multi-column layouts and forms
1.4.3 Contrast (Minimum)Text over backgrounds and images meets 4.5:1 - the criterion PDF/UA does not check
1.4.5 Images of TextNo scanned-image pages: OCR + tagging so text is real and selectable
2.4.2 Page TitledDocument Title set in metadata and displayed instead of the file name
2.4.5 Multiple WaysBookmarks mirroring the heading structure in long documents
3.1.1 Language of PageThe /Lang entry set on the document, and on passages that switch language
3.3.2 Labels or InstructionsForm fields with tooltips carrying the visible label
4.1.2 Name, Role, ValueInteractive form fields exposing name, role and state through tags and the form dictionary

Matterhorn: how PDF/UA conformance is tested

The Matterhorn Protocol, published by the PDF Association, translates PDF/UA into a concrete test plan: 31 checkpoints broken into failure conditions, each marked as machine-verifiable or requiring human judgement. That split mirrors web testing: software can prove a tag tree exists, fonts map to Unicode and metadata is set; only a person can confirm the alt text is meaningful and the reading order makes sense.

This is exactly how the EqualWeb PDF checker works: 80+ automated checks across 8 categories (structure, tagging, reading order, alternate text, tables, forms, metadata, fonts) validate against both PDF/UA-1 and PDF/UA-2, and the report separates what is machine-proven from what needs expert review. Conformance is reported as binary - a document passes or it does not - because that is how PDF/UA defines it.

Making a PDF library accessible: the workflow

  1. Inventory and triage. Most organisations discover thousands of PDFs; not all matter equally. Prioritize documents that are part of active services - forms, statements, current guides.
  2. Automated check. Run the collection through the checker (single files up to 400 pages / 100MB, batches up to 1,000 documents) to map failures by category.
  3. Remediate. AI auto-remediation rebuilds tagging, reading order and metadata at scale; certified experts handle complex tables, forms and judgement calls.
  4. Fix at the source. Update the Word/InDesign templates that generate the documents, so next month's output is born accessible instead of remediated again.
  5. Verify and keep evidence. Re-check remediated files and retain the reports - the same evidence logic as a VPAT on the web side.

Frequently asked questions

Does WCAG apply to PDF files?

Yes. The W3C publishes PDF-specific techniques for meeting WCAG, and the laws that reference WCAG - EN 301 549 (Chapter 10), Section 508, the ADA - apply it to published documents, not only web pages.

What is the difference between WCAG and PDF/UA?

WCAG defines accessible outcomes for any content; PDF/UA (ISO 14289) defines how a PDF file must be built internally - tags, reading order, Unicode, metadata. PDF/UA does not cover visual requirements like contrast, so documents are tested against both.

What is a tagged PDF?

A PDF with an internal structure tree (tags) that tells assistive technology what each element is - headings, paragraphs, lists, tables, figures - and in what order to read them. Without tags a screen reader gets an unstructured stream or nothing.

Can a scanned PDF be accessible?

Not as-is: a scan is a picture of text and fails WCAG 1.4.5 and 1.1.1. It becomes accessible through OCR (real text) plus tagging - a standard step in remediation.

Is PDF/UA-1 enough, or do I need PDF/UA-2?

PDF/UA-1 (ISO 14289-1) covers PDF 1.7 files; PDF/UA-2 (ISO 14289-2) is the current part for PDF 2.0. Validate against both - the EqualWeb checker tests both, with UA-2 as the engine default.

How do I check if a PDF is accessible?

Run it through an automated PDF accessibility checker for the machine-verifiable failures (tags, metadata, fonts, alt-text presence), then have the judgement items - reading order, alt-text quality, table logic - reviewed by a person. The EqualWeb checker runs 80+ checks across 8 categories, free.

Sources
  • W3C WAI, PDF Techniques for WCAG - w3.org/WAI
  • ISO, ISO 14289 (PDF/UA) parts 1 and 2 - iso.org
  • PDF Association, Matterhorn Protocol - pdfa.org
  • ETSI, EN 301 549, Chapter 10: Non-web documents - etsi.org

Last updated . Reviewed by the EqualWeb accessibility team.

Check any PDF against WCAG and PDF/UA - free

80+ automated checks across 8 categories, validated against PDF/UA-1 and PDF/UA-2, with AI remediation and certified experts behind it. No credit card required.

Book a meeting Start Free Trial