Paste your palette. See it the way people with each type of color blindness do — and get the exact color pairs that become hard to tell apart. Scored with a published, open-access method, entirely in your browser.
📷 Drop a screenshot or mockup — or click to upload, or paste — and we'll pull out its main colors
…or add & edit colors by hand
What this is — and isn't. This checks one specific thing: whether your colors stay distinguishable under simulated color-vision deficiency. It is not a legal accessibility audit and does not certify ADA, Section 508, WCAG, or EU Accessibility Act compliance — those cover far more than color. It simulates a model of color blindness on your uncalibrated screen, not any single person's vision. Use it to catch color conflicts early; use a full review for compliance.
Method. Simulation uses Machado, Oliveira & Fernandes (2009) severity-1 matrices in linear RGB; perceptual difference uses CIEDE2000. A pair is flagged when it is clearly distinct to normal vision but collapses under a CVD simulation. This is the same color-vision science behind the OpticQuiz published paper — doi.org/10.5281/zenodo.21310578. The engine is open source: github.com/zengineco/opticquiz.com.
What a colorblind-safe palette is
A colorblind-safe palette is a set of colors that stay distinguishable to people with color vision deficiency. Red-green deficiency alone affects roughly 1 in 12 men, so any design that uses red-versus-green to carry meaning — chart series, status colors, maps, wayfinding — risks becoming ambiguous. This checker simulates all three deficiency types (protan, deutan, tritan) and reports the specific pairs that need attention, so you can fix them before they ship.
How the checker works
For every pair of colors that is clearly distinct to normal vision, the tool simulates how the pair looks under each type of color blindness and re-measures the perceptual gap with CIEDE2000. If that gap collapses, the pair is flagged — as a risk (hard to tell apart) or severe (effectively the same). It never invents a grade: you see the exact before-and-after difference for every conflict.
What the result means
A pass means no clearly-distinct pair collapses under simulation — a good sign your palette carries its meaning by more than just hue. A fail lists exactly which pairs to fix: nudge one color's lightness, add a non-color cue (labels, patterns, icons), or swap to a hue that survives the confusion axis. Re-run until it passes. It is a screening aid, not a guarantee, and it never claims legal compliance.
A set of colors that stay distinguishable to people with color vision deficiency. Because red-green deficiency is common (~8% of men), palettes that rely on red-vs-green can become ambiguous; a colorblind-safe palette keeps every meaningful pair distinct after simulating protan, deutan and tritan vision.
Does passing mean my design is ADA compliant?
No. This checks color distinguishability only. It is not a legal audit and does not certify ADA, Section 508, WCAG, or EU Accessibility Act compliance, which cover much more than color. Use it to catch color conflicts; use a full review for compliance.
How accurate is the simulation?
It uses a peer-reviewed model (Machado et al., 2009) and a standard perceptual metric (CIEDE2000). It reliably catches classic red-green and blue-yellow conflicts, but it models a diverse population on an uncalibrated screen — an approximation, not any one person's vision.
