Testing methodology

How Allie produces its results, and how to read them.

The engine

Allie uses axe-core, a widely used open-source accessibility testing engine. Every report states the axe-core version that produced it. Allie loads the page in Chromium, waits for it to finish loading, then runs axe-core on the rendered page.

Which rules run

Allie enables axe-core rules tagged for these WCAG levels:

Together this tests against WCAG 2.2 A and AA. Experimental rules and general best-practice rules that are not tied to a WCAG criterion are not included, and Level AAA is not tested.

Reading the results

Fail (an issue)
axe-core is confident that an element breaks a rule. Each failing element counts as one issue. Severity (critical, serious, moderate, minor) is the impact rating given by axe-core for that rule.
Pass
A rule applied to the page and found no problems on the elements it checked.
Needs review
axe-core could not decide automatically, for example when text sits over an image or gradient and contrast cannot be measured reliably. These are not counted as issues. A person should check them by hand, so they appear as a count in the report.
Not applicable
The page contains nothing the rule applies to, such as a video rule on a page with no video.

Scan conditions

What automated testing cannot do

Automated tools cannot judge whether alt text is meaningful, whether content is understandable, whether the focus order is logical, whether captions are accurate or whether a task is usable with a screen reader. Automated testing covers only part of WCAG, so a page with zero automated failures is not necessarily accessible or conformant.

Allie's suggested fixes are generic starting points. A developer should review them in context.

Not legal advice

Allie reports are technical test results. They do not certify conformance with WCAG, the ADA or any other law, and do not protect against claims. For legal questions talk to a qualified professional. Questions about our method? hello@a11ie.io.