WCAG 2.1 AA
violations
- Level A
- Level AA
- Criteria passed
- Compliance
The regulatory baseline. EN 301 549 and the EU Web Accessibility Directive still reference WCAG 2.1.
Paste a web address. WARD fetches the homepage once, evaluates it under WCAG 2.1 and WCAG 2.2 side by side, explains each violation, places the result in the five-country census distribution, and turns the findings into remediation tasks.
violations
The regulatory baseline. EN 301 549 and the EU Web Accessibility Directive still reference WCAG 2.1.
violations
Each violation type becomes one task unit, as in the crowdsourcing framework presented at UD26. Non-technical units need a browser and the site's content editor; technical units need someone who can change templates or code. Level A first, then by recurrence.
Between June and July 2026 WARD audited the homepages of 161,220 civil society organisations with a recorded website in the registers of the United Kingdom, Australia, the United States, New Zealand and Ireland. 135,604 could be reached. The distribution shown above comes from those records, released as an open dataset (Zenodo DOI 10.5281/zenodo.23529469). Static analysis detects a subset of real violations, so counts are floors; the browser-based validation in the dataset quantifies the floor per criterion.
WARD is a static-analysis checker. It fetches the HTML of one page, does not run scripts, and evaluates 18 WCAG 2.1 criteria plus the four criteria WCAG 2.2 introduces. It does not test colour contrast or anything that requires rendering. A clean result here means the page passes the machine-checkable floor, not the whole standard.
Code and data: cso-web-accessibility repository on GitHub. Contact: Muhammad Sohaib Ayub, Institute for Data and AI Research, University of Galway.