Guide
PDF/UA vs WCAG: which one applies to your documents
One is a format standard a validator can test. The other is what your legal obligation cites. They overlap heavily and neither one contains the other.
About 10 minutes · quick
The short answer
WCAG 2.1 AA is what your legal obligation almost certainly cites. It is technology-neutral: written for web content, applied to documents by analogy, and it says what must be true without saying how a PDF should express it.
PDF/UA-1 (ISO 14289-1) is a format standard. It says exactly which PDF objects must exist and how they must relate. It is the only one of the two that a validator can test mechanically, because it talks about the file rather than about the experience.
They are not alternatives. In practice: conform to WCAG because that is the obligation, validate against PDF/UA because that is the test.
What each actually says
WCAG 2.1 Level AA gives you criteria like 1.3.1 Info and Relationships: information conveyed through presentation must be programmatically determinable. True of a web page, a PDF, an EPUB. It does not tell you that the heading must be an H2 structure element whose parent chain is valid, because it is not about PDF.
PDF/UA-1 does. Its clause 7.4 is specifically about headings in a PDF: which element types, how levels may progress, that a document must not mix H with H1–H6. Its clause 7.1 is about the structure tree existing and every painting operator being in it or artifacted. Its clause 5 is about the XMP identifier.
So PDF/UA is more specific, more testable, and much narrower. WCAG covers things PDF/UA does not mention at all.
What only WCAG covers
These are real failures in real documents, and no PDF/UA validator will ever report them:
- Colour contrast (1.4.3). Nothing in PDF/UA is about contrast. A document can validate perfectly with 2:1 grey text.
- Use of colour (1.4.1). A chart whose only legend is colour.
- Link purpose (2.4.4). PDF/UA requires a link annotation to have a description. WCAG requires the description to identify the purpose. Click here satisfies one and not the other.
- Meaningful sequence (1.3.2), beyond the structural. PDF/UA requires a reading order to exist; WCAG requires it to be the correct one, which is a judgement no validator makes.
- Text alternatives that convey purpose (1.1.1). PDF/UA requires
/Altto be present and non-empty. Whether it says anything useful is WCAG’s question and no machine’s.
What only PDF/UA covers
These are PDF-specific and WCAG has no language for them:
- The role map. Custom tag names must resolve to standard types, and standard types must never be remapped (02-001, 02-004).
- The PDF/UA identifier in the XMP metadata (06-002) — a machine-readable claim with no experiential effect at all.
DisplayDocTitle(07-001), which decides whether the window title is the document’s name or the filename.- Artifact and real-content nesting rules, character-to-Unicode mapping requirements, font embedding requirements, the
Suspectsflag, optional-content dictionary naming.
Several of these have no effect on any screen reader. They are conformance machinery — the apparatus that makes the claim testable.
Where they overlap, which is most of it
Structure, order, alternatives and language are the substance of both. The Matterhorn Protocol’s failure conditions carry explicit WCAG mappings for exactly this reason: 09-001 (reading order) maps to 1.3.2, 14-003 (heading levels) to 1.3.1, 13-001 (untagged graphics) to 1.1.1, 11-001 (document language) to 3.1.1, 15-002 (table headers) to 1.3.1.
Work through PDF/UA-1 properly and the overlap is done. What remains for WCAG is contrast, link purpose, and whether the alternatives you wrote actually say anything — which is to say, the part that was always going to need a person.
Which to cite, by situation
A public body in the EU. EN 301 549 clause 10 is the harmonised standard, and it points at WCAG 2.1 AA for non-web documents. Cite WCAG; use PDF/UA-1 as the test. EN 301 549.
A US federal agency or contractor. Section 508’s refresh incorporates WCAG 2.0 AA by reference; procurement commonly asks for 2.1. Section 508.
A US state or local government. ADA Title II’s rule sets WCAG 2.1 AA, with deadlines of 26 April 2027 and 26 April 2028 depending on the population served. ADA Title II.
Israel. IS 5568 adopts WCAG and adds Hebrew-specific requirements, which is where right-to-left reading order stops being a nicety. IS 5568.
A private company selling into the EU. The European Accessibility Act has applied since June 2025 and covers documents that are part of a covered service. EAA.
In every one of those the obligation is WCAG and the testable proxy is PDF/UA-1. Nobody is asking you to pick.
What PDF/UA-2 changes
ISO 14289-2 was published in 2024, built on PDF 2.0. It tightens several things PDF/UA-1 left ambiguous and adds a namespace mechanism for structure types. For almost everyone it changes nothing yet: the tooling is thin, the validators are young, and procurement still asks for PDF/UA-1. What changes, and when it will matter.
How to actually check either one
PDF/UA-1: veraPDF is the reference implementation and is free. Taggart checks all 87 machine-decidable Matterhorn conditions free and unlimited, and tells you which of the remaining 47 need a person.
WCAG 2.1 AA for a document: the machine-checkable overlap is covered by the above. Contrast needs a contrast check. Link purpose, alt-text quality and reading-order correctness need a screen-reader pass. There is no tool for the last three and there will not be one.
Check a PDF free → · Every failure condition, mapped to WCAG →
Frequently asked questions
If I do one, do I get the other?
Mostly, in one direction. A document that genuinely meets PDF/UA-1 meets most of WCAG 2.1 AA's document-relevant criteria, because PDF/UA is largely WCAG's structural requirements expressed in PDF objects. The reverse is weaker: WCAG says nothing about role maps, the PDF/UA identifier, or several PDF-specific requirements, so a WCAG-shaped effort leaves PDF/UA failures behind.
Which do I put in a contract?
Both, with different verbs. WCAG 2.1 AA is the obligation, because that is what the law cites. PDF/UA-1 is the acceptance test, because it is the one a validator can actually run. "Conforms to WCAG 2.1 AA; validates clean against PDF/UA-1 (ISO 14289-1); accompanied by a remediation report recording human decisions" is a clause that can be checked.
What about the Matterhorn Protocol?
It is not a third standard. It is the PDF Association's breakdown of PDF/UA-1 into 31 checkpoints and 136 specific failure conditions, 87 of which a machine can test. It is how you check PDF/UA-1, not something to conform to separately.
Does WCAG 2.2 change anything for PDFs?
Very little. The criteria added in 2.2 are overwhelmingly about interactive web interfaces — focus appearance, dragging movements, consistent help. For a static document almost nothing changes, which is why most document policy still cites 2.1 AA.
Is PDF/UA legally required anywhere?
Rarely by name. Most law cites WCAG, or cites a standard that cites WCAG — EN 301 549 in Europe, Section 508's refresh in the US. PDF/UA appears in procurement specifications far more often than in statute, because buyers want something testable.
Related failure conditions
- 09-001 Tags are not in logical reading order The order of elements in the structure tree does not match the order a human would read the page. This is the single most common substantive defect in tagged PDFs and the one Acrobat's own tooling handles worst.
- 14-003 Numbered heading levels in descending sequence are skipped The heading hierarchy jumps a level going down — H1 straight to H3. Going back up any number of levels is fine; going down must be one level at a time.
- 13-001 Graphics objects are not tagged with a Figure tag An image or vector drawing that carries meaning is not inside a Figure element. Either it is untagged, or it has been swept into a paragraph.
- 11-001 Natural language for text in page content cannot be determined No /Lang is in effect for some text: neither on the catalog, nor on an enclosing structure element, nor on a marked-content sequence.
- 15-002 A column has a header cell, but that header cell is not tagged as a header The top row labels the columns but is tagged as ordinary data cells.
Check your own file. Taggart's validator is free and unlimited — no page cap, no watermark, no account.
Last updated .