IS YOUR PALETTE
COLORBLIND-SAFE?
The default palettes in your charting library were mostly not designed for color-vision deficiency. Here is every common one, checked by the same published method behind our tools — with the exact conflicts, and a generated safe version of each one that fails.
CATEGORICAL PALETTES
Discrete categories — every color must stay tellable-apart from every other. Flagged when a pair that's distinct to normal vision collapses under simulated protan, deutan or tritan vision (CIEDE2000). Failing palettes show an auto-generated colorblind-safe fix.
COLORMAPS
Sequential and diverging ramps are read by lightness, not by telling swatches apart — so the right test is different: does perceived lightness stay monotonic under CVD? A big reversal means two data values map to the same brightness.
opticquiz-cvd — the same engine as the checker, method published at DOI 10.5281/zenodo.21310578. Nothing is hand-entered. As a cross-check, a second independent model (Brettel 1997) agrees with the primary model (Machado 2009) on of the categorical palettes; the lone difference is a single borderline tritan pair in Okabe–Ito.
Paste it into the checker → to see its conflicts and get a fix, or check text contrast →.
QUESTIONS
Yes, used as intended. viridis is a sequential colormap read by lightness, and its lightness stays monotonic under all three CVD types (zero reversal here) — so order survives. Don't use it as a set of distinct categorical colors.
No. jet's lightness is strongly non-monotonic under simulation (~37-point L* reversal), so different values collapse to the same brightness. That's the long-standing case against jet.
For categories: Okabe–Ito or Tableau Colorblind 10 (both pass here). For continuous data: viridis and its siblings. Or paste your brand colors into the checker and use the generated fix.