Khroma

Khroma is the color tool that learns your taste instead of debating it — a small, focused application of machine learning that solves a genuinely recurring problem: choosing color palettes you actually like. The onboarding is the idea: you browse fifty pairs of colors, picking the combinations that appeal, and Khroma trains a personal algorithm on your choices. From then on it generates infinite palettes tuned to your demonstrated preferences — not a generic 'trending colors' list, but palettes that feel like yours, ready as CSS, hex, RGB, and Tailwind-friendly formats.

The practical depth exceeds the cute onboarding. Every generated palette arrives pre-checked for accessibility: contrast ratios against white and black text shown on the spot, so WCAG compliance becomes a property of the palette rather than a later audit. Searchable palettes by hue, browseable grids for exploration, and the ability to favorite and export make it usable in real design workflows. In our design-team testing, it earned a permanent tab for brand exploration phases — producing starting points that survived into actual mockups far more often than stock palette sites.

Honest limits: the model reflects what you picked, so rushed or inconsistent onboarding yields muddled recommendations; it generates pairings, not full design systems or semantic token structures; and there is no collaboration layer — it is a personal instrument. Free, with no meaningful paywall — remarkable for the utility. For designers, developers, and anyone who has ever lost an hour to a palette picker, Khroma is thirty minutes of training that pays off for years. Small, free, and genuinely smart.