WCAG 2.2 AA checks for generated PDFs

Evidence, not a certification. This page publishes automated checks on the listed source HTML and generated PDFs, with tool versions, raw reports, and hashes.

WCAG is a property of the final document, not just the API. Fork PDF runs automated checks on generated PDFs and the source HTML used to produce them. The samples below cover headings, paragraphs, lists, tables, links, images with alt text, and a multi-page document.

HTML to PDF mapping

The PDF is generated from the source HTML. Alt text, reading order, and document structure come from that source.

Source HTML	PDF output	Preserved
`h1`–`h6`	`H1`–`H6`	Text, heading level, and page order.
`p` and text nodes	`P` / `Span` tagged content	Text content and source order.
`ul`, `ol`, `li`	`L`, `LI`, `LBody`	List item text and order.
`table`, `th`, `td`	`Table`, `TR`, `TH`, `TD`	Row, header, and cell structure with header and cell text.
`a href`	`Link` plus link annotation	Link text, URI, and annotation bounds.
`img alt` and `svg role="img" aria-label`	`Figure` with `Alt`	Alt text is written verbatim from the source HTML.
Text runs	Embedded fonts with `ToUnicode`	Extractable words and copy/paste text.

What we test

Travel itinerary: four-page PDF and its rendered source HTML.
Invoice: headings, line items, table cells, and totals.
Link and image: a link annotation plus an image with alt text.
Multi-link: three links across two pages.

Evidence version and provenance

Evidence version: 2026-10-31. Results apply to the exact source and PDF files listed below. The manifest records tool versions, file sizes, and hashes. hashes.txt verifies the evidence set with sha256sum -c hashes.txt.

Sample	Source HTML	PDF	PDF SHA-256
Travel itinerary	itinerary.html	itinerary-wcag.pdf	`bf89547e495b5cc608460d453f09503c85d00295496fb2555a5a55805fe4c589`
Invoice	invoice.html	invoice-wcag.pdf	`eb5eaeb558eb279055c35365b450f75b64f6c7c6660db42fbc9c6f348d0a4a4e`
Link + image	links.html	links-wcag.pdf	`05b4838c1b14402b7f7aca5dab5d5512d744bd148d521db7251dc3affd5cde84`
Multi-link	multilink.html	multilink-wcag.pdf	`24973e6fa1e08c72f5c32f55da313de63b3c4de2c87d9dfd25beeeef46959555`

Automated results

Sample	axe-core HTML	axes4 WCAG	Notes
Travel itinerary	0 violations, 1 incomplete	0 errors / 0 warnings	Incomplete is SVG color contrast; calculated contrast passes. PDF.js extracts whole words.
Invoice	0 violations, 0 incomplete	0 errors / 0 warnings	Headings, table, and totals pass
Link + image	0 violations, 0 incomplete	0 errors / 0 warnings	Link annotation and image alt pass
Multi-link	Not audited as HTML	0 errors / 0 warnings	Three link annotations across two pages

axe-core 4.10.2 and Chromium 151.0.7880.0 ran at 1280×900 and 320×900. The axes4 WCAG2CheckSet ran on the generated PDFs; job IDs and raw reports are included in the evidence bundle. PDF.js 4.10.38 produced the text-extraction evidence for the itinerary. PDF/UA evidence is published separately and is not used for this WCAG result.

Known limitations

The travel itinerary is a fixed A4 layout. WCAG 1.4.10 reflow is not demonstrated at 320 CSS px.
Avoid CSS letter-spacing on text that must be extracted. Position-based PDF readers can insert spaces between glyphs even when the visual text is correct.

Download the evidence

File	Details
wcag-2.2-aa-proof.zip	Complete evidence bundle: sample PDFs, source HTML, raw tool reports, manifest, and hashes.
itinerary-wcag.pdf	Four-page travel itinerary.
invoice-wcag.pdf	Invoice with headings and a table.
links-wcag.pdf	Link and image sample.
multilink-wcag.pdf	Three links across two pages.
manifest.json	Evidence version, tool versions, sample hashes, result counts, and file hashes.
axe-results.json	Full axe-core results, including incomplete-node details, at both widths.
axes4-summary.json	axes4 counts and job IDs for all four PDFs.
axes4 raw reports	itinerary, invoice, links, and multilink.
contrast-calculations.txt	Color pairs, ratios, and the calculation method for the SVG map.
pdfjs-text-extraction.txt	PDF.js text extraction from the fixed itinerary sample.
hashes.txt	SHA-256 values for the evidence files. Verify with `sha256sum -c hashes.txt`.
zip-hash.txt	SHA-256 of the evidence bundle. Verify with `sha256sum -c zip-hash.txt`.

Reproduce

Download the evidence bundle and run sha256sum -c hashes.txt.
Open the source HTML from the provenance table and run axe-core 4.10.2 at 1280×900 and 320×900. Compare the violations and incomplete results with axe-results.json.
Upload the four PDFs to axes4 WCAG2CheckSet. Compare the errors, warnings, and job IDs with axes4-summary.json and the raw reports. axes4 access is required.
Run PDF.js 4.10.38 text extraction on itinerary-wcag.pdf and compare it with pdfjs-text-extraction.txt.

Evidence history

2026-10-31 — Evidence bundle refreshed with source HTML, full axe results, raw axes4 reports, an evidence manifest, and hashes.

This page

Automated page check: axe-core 4.10.2 reported 0 violations and 0 incomplete results at 1280×900 and 320×900 on 2026-10-31. page-axe-results.json

Report an evidence error

If a file, hash, or result on this page is wrong, contact us with the page URL and evidence version 2026-10-31.

Need a full accessibility audit?

Use this page as a starting point, not as legal advice or a conformance certificate. This page reports automated checks only. Contact us if you need help scoping manual testing or remediation.