Graphics have clear contrast against background
What this criterion checks
This criterion maps to 1.4.11. Non-text contrast (AA). A detailed plain-language description is maintained in the WebLens criteria catalogue; the WCAG sub-test breakdown below describes exactly what the agent checks and what it does not.
How it's assessed
WebLens walks each user journey, captures a screenshot at every step, and an agent reasons over that evidence. See the how WebLens works for the full pipeline and the deterministic checks layered in for accessibility.
Sub-test coverage
1.4.11. Non-text contrast (AA) is broken into 4 sub-tests. Coverage is honest about where the agent is reliable, where it only partially checks the requirement, and where a human reviewer is still needed.
1 full3 partial
- Interactive controls are visually identifiable — the visual cues that help a user spot a control on the page (such as a border around a text field, the shape of a checkbox, or a dropdown arrow) have at least 3:1 contrast against the surrounding areaFull
Our tool measures the contrast of borders, outlines, and background colours of form fields, dropdowns, and similar controls against the surrounding page. Controls clearly identified by their text label or icon do not need a separate visible border — the tool recognises this and does not flag them. Controls using the browser's default appearance (not styled by the website) are automatically excluded, as the criterion exempts them. A person should verify any flagged control where the visual indicator is created using techniques the tool cannot inspect, such as decorative effects layered over the control.
- Control states are visually clear — when a user focuses, checks, or selects a control, the visual change that shows the new state (such as a focus outline, a check mark, or a toggle position) has at least 3:1 contrastPartial
Our tool tests focus indicators and checked/selected states by activating each control and comparing the visual appearance before and after. It detects whether a visible indicator appears and whether it has sufficient contrast. However, some websites only show focus indicators when using a keyboard (not when clicking), and the tool activates controls by simulating a click — this can cause it to flag controls that do have a visible keyboard-only focus ring. A person should confirm focus-related findings by tabbing through the page with a keyboard.
- Meaningful parts of graphics have at least 3:1 contrast — icons, chart segments, diagram elements, and other visual parts that a user needs to see in order to understand the content are distinguishable from the area around themPartial
Our tool detects common graphical elements such as icons and simple graphics and measures their contrast against the surrounding area. For more complex graphics like charts, diagrams, and data visualisations, deciding which parts are needed to understand the content often requires human judgment. A person should review multi-colour visualisations to verify that each meaningful segment is distinguishable from its neighbours.
- Exceptions are correctly applied — disabled controls, controls with unmodified browser styling, and graphics where the specific appearance is essential (such as logos, photographs, and flags) are exempt from the 3:1 requirementPartial
Our tool excludes disabled controls and recognises unmodified browser-styled checkboxes, radio buttons, and sliders as exempt. For other control types (such as buttons and text fields), the tool does not yet distinguish between browser-default and website-styled appearances. For graphics, the tool uses visual analysis to assess whether a presentation might be essential, but determining whether a specific image truly cannot be presented differently (such as a national flag or a medical diagram) requires human judgment.
How reliable it is
This criterion performs well against expert review. Treat its findings as strong leads and confirm before remediating. See Accuracy & confidence for how the confidence tags are set.