Free Accessibility Audit: What It Finds vs What It Misses

Every accessibility tool offers a free scan. Including ours. But what does a free scan actually tell you?

The honest answer: a lot, but not everything. Understanding the boundary between what automation catches and what it misses is the most important thing you can know about accessibility testing.

What automated scanning catches well

Automated tools excel at detecting issues that have clear, objective rules:

Missing alt text. If an image has no alt attribute, any scanner will flag it. This is the most commonly detected issue across all websites.

Colour contrast failures. Tools can compute the contrast ratio between foreground text and background colours and compare against the WCAG 4.5:1 ratio for normal text.

Missing form labels. Input fields without associated labels are straightforward to detect in the DOM.

ARIA errors. Invalid ARIA roles, empty aria-labels, broken aria-labelledby references, and redundant roles on semantic elements.

Structural HTML problems. Skipped heading levels, missing landmarks, multiple h1 elements, tables without headers.

Missing language attribute. The html element needs a lang attribute for screen readers to use the correct voice.

A good multi-engine scanner (running axe-core, HTML_CodeSniffer, and custom rules) will catch all of these reliably.

What automation catches partially

Some issues can be detected but not fully evaluated:

Alt text quality. A scanner can tell you alt text exists and flag obvious problems (filename as alt text, alt text that's too short, duplicate alt text). But it cannot tell you whether the alt text is actually meaningful and descriptive.

Link text. Scanners can flag "click here" and "read more" as vague link text. But they can't evaluate whether specific link text adequately describes the destination.

Focus visibility. Keyboard testing engines (like Lumi's) can detect missing focus indicators and focus trapped behind sticky elements. But evaluating whether a custom focus style is sufficiently visible requires visual judgment.

What automation misses entirely

These require a human expert:

Screen reader experience. Does the page make sense when read linearly? Are decorative images properly hidden? Do dynamic updates get announced? Is the reading order logical?

Cognitive accessibility. Is the language clear? Are instructions understandable? Is the layout consistent? Are error messages helpful?

Interaction quality. Can a user complete a multi-step form? Does a modal behave correctly with a screen reader? Does a custom dropdown work with keyboard and assistive technology?

Content meaningfulness. Is this heading accurate? Does this table make sense? Is this error message actionable?

Context-dependent decisions. Whether a particular image needs alt text (functional images always do, decorative images shouldn't have it).

The practical takeaway

Use free automated scanning to:

• Establish your baseline score
• Find the obvious issues you can fix immediately
• Monitor for regressions after changes
• Prioritise by severity

Invest in a manual audit when:

• You need actual WCAG compliance (not just a score)
• Your site has complex interactive features (forms, e-commerce checkout, dashboards)
• You're facing a legal deadline (EAA, ADA)
• You want to genuinely serve users with disabilities, not just check a box

Lumi gives you the most thorough automated scan available - five engines, 350+ rules, including keyboard testing that most tools skip entirely. It's free to try at lumi.livana.io.

For the manual audit side, Livana's team tests with real assistive technology across all major screen readers. Learn more at livana.io.