A11yDen

Colour Blindness Simulator

Enter your colours and see them as viewers with eight types of colour vision deficiency would — then check the part that actually breaks designs: whether pairs you use to mean different things (error vs success, two chart series, two teams) can still be told apart by colour alone.

Simulate your colours

How the simulation works

The three dichromacy simulations (protanopia, deuteranopia, tritanopia) use the colour-transformation matrices published by Machado, Oliveira & Fernandes (2009), the standard physiologically-based model, applied to your colours exactly as published. The “-anomaly” (colour-weak) types are interpolated at roughly 60% of the way from normal vision to the matching dichromacy — a documented approximation, labelled as one. Achromatopsia keeps only the luminance of each colour. Pair distinguishability uses the same contrast engine as the Contrast Checker, run on the simulated colours.

Honest limits: a simulation is one model of vision, not a photograph of anyone's eyes. Real colour vision varies by person and by screen. Treat a flagged pair as “redesign this so it does not depend on colour alone”, and an unflagged pair as “no problem found by this model” — not as a guarantee.

Criterion this tool informs
SCCriterionLevel
1.4.1Use of ColorA

What this tool cannot check

Whether your design actually relies on colour alone — that depends on your labels, icons, patterns and layout, which only a person reviewing the real page can judge. It also cannot reproduce any individual's exact vision, and it says nothing about the contrast of text, which is the Contrast Checker's job.

Questions

How accurate is the simulation?

It implements a peer-reviewed model (Machado et al., 2009) with its published dichromacy values. That is the same class of model used by established tools, and it is still an approximation: severity differs between people, and anomalous (colour-weak) vision is shown at an interpolated severity rather than a measured one. Use it to catch designs that depend on colour alone — the failure it finds is real even when the exact simulated shade is approximate.

Why 1.2:1 as the “hard to tell apart” line?

It is this tool's stated rule of thumb, not a WCAG threshold: below roughly 1.2:1 simulated contrast, two colours carry almost no visible difference for that viewer. WCAG itself sets no ratio for colour pairs — SC 1.4.1 simply forbids colour from being the only means of conveying information.

Related tools

All tools

All ten tools are listed on the homepage.