Overview
AIWebPageSEO combines custom accessibility checks with selected checks from axe-core, version 4.10.2. The axe-core engine is served from a local JavaScript bundle and runs selected checks in an isolated frame containing a sanitized representation of the page being evaluated.
The list below identifies 14 selected axe-core rules carrying the ACT tag in the installed axe-core 4.10.2 rule metadata. The tag indicates an ACT association within axe-core; it is not a W3C ACT Rule identifier, nor is it a claim that every aspect of a corresponding published ACT Rule has been implemented or verified.
WCAG 2.2 Level A and AA assessment coverage
The Accessibility Audit maintains a criterion-by-criterion coverage list for all 55 WCAG 2.2 Level A and AA success criteria. The production implementation classifies 4 criteria as fully automated, 31 as partially automated, and 20 as requiring human evaluation. These classifications describe how the tool evaluates each criterion; they are not a claim that every criterion has been conclusively passed.
For each criterion, the report identifies one of five outcomes: Failed, Warning, No problems found, Needs a person to check, or Not checked this time. Partially automated criteria can also display specific guidance about what still needs human verification, such as the accuracy of image alternatives, completeness of captions, or functionality beyond the keyboard interactions tested.
A “No problems found” result means the checks available for that criterion ran without a mapped failure or warning. For a partially automated criterion, it does not certify that all aspects of the success criterion have been conclusively verified.
Rendered-browser and interaction testing
AIWebPageSEO goes beyond static markup inspection. Its server-side rendered accessibility analysis uses a headless browser to evaluate pages as rendered and includes keyboard navigation, focus behaviour and potential keyboard traps; rendered colour and non-text contrast; zoom and reflow; portrait and landscape orientation; text spacing; moving and flashing content; and selected hover or focus interactions.
The audit also examines selected inner pages and compares site-level features such as consistent navigation and consistent help placement. Test scope is bounded by what can be loaded and exercised during a run; the report identifies checks that did not run and criteria needing further human evaluation.
The assessment combines these rendered-page tests with custom accessibility checks and the locally hosted axe-core 4.10.2 integration described below.
Selected ACT-tagged axe-core rules
These rule IDs appear in the production runner's explicitly selected rule lists and carry an ACT tag in the installed engine. A rule may yield no result if the relevant elements or conditions are absent; execution may also be limited by page rendering, sanitization, or other audit conditions.
| axe-core rule ID | Area evaluated | Runner category |
|---|---|---|
area-alt | Image-map area alternatives | WCAG-oriented |
aria-command-name | Accessible names for ARIA command controls | WCAG-oriented |
aria-input-field-name | Accessible names for ARIA input controls | WCAG-oriented |
aria-toggle-field-name | Accessible names for ARIA toggle controls | WCAG-oriented |
autocomplete-valid | Valid autocomplete tokens | WCAG-oriented |
avoid-inline-spacing | Text spacing restrictions | WCAG-oriented |
document-title | Document title | WCAG-oriented |
html-xml-lang-mismatch | Consistency of HTML and XML language declarations | WCAG-oriented |
input-button-name | Accessible names for input buttons | WCAG-oriented |
meta-viewport | Viewport zoom and scaling restrictions | WCAG-oriented |
presentation-role-conflict | Conflicts involving presentational roles | Best-practice note |
role-img-alt | Accessible names for elements with image roles | WCAG-oriented |
svg-img-alt | Accessible names for SVG images | WCAG-oriented |
valid-lang | Validity of language values | WCAG-oriented |
How these checks are run
The rendered-page audit loads a locally hosted axe-core bundle (public/a11y-axe/axe.min.js) and runner (public/a11y-axe/run.js). The runner selects named rules, filters out rule IDs not present in the loaded engine, and invokes axe.run with runOnly set to those rules. It collects reported violations, including rule ID, impact, descriptive guidance, and examples of affected elements.
The runner also selects additional rules without an ACT tag. Best-practice results are identified separately from WCAG-oriented results. AIWebPageSEO's own checks provide further coverage beyond this selected axe-core rule set.
Interpretation and limitations
- The
ACTmetadata tag is an axe-core classification. This page does not claim one-to-one correspondence or verified equivalence with individual published W3C ACT Rules. - Only the explicitly selected axe-core rules are run by this integration; other rules present in axe-core 4.10.2 are not implied to be enabled.
- The audit evaluates a sanitized representation of page content. Dynamic states, embedded content, user interactions, or content removed for isolation may require separate testing.
- The tool reports coverage across all 55 WCAG 2.2 A/AA criteria, including criteria for which it explicitly requests human evaluation. Its automated findings alone do not constitute a final conformance determination; the report identifies the remaining review needed rather than treating an untested aspect as a confirmed pass.
- Actual findings depend on the content and conditions of each audited page; the presence of a rule in the configuration does not guarantee it produced a result in every audit.