Upload an image and get its most frequent colors as hex swatches, ranked by how much of the image they actually cover — a histogram over the real pixels, not a guess, with each swatch copyable in one click.
Analyzed locally, the image never uploads
Drop it in or click to choose a file.
From 2 up to 12 swatches.
Click any swatch to copy its hex value.
Every pixel is rounded to one of 8 levels per RGB channel — a 512-cube grid — and pixels landing in the same cube are counted together and averaged. The most populated cubes become the swatches. This is a histogram over real pixel data, explainable and honest about what it does, not a perceptual clustering algorithm claiming to find "the" definitive palette — different images can genuinely have more than one reasonable answer.
The extraction ranks by actual pixel frequency in a 512-color-cube histogram, not by visual prominence or perceptual importance — a small but frequent color can outrank a large but visually striking one that only appears in a few distinct shades.
No — pixels with low alpha are excluded, so a transparent background does not show up as a dominant color.
No. The analysis happens entirely in your browser using the Canvas API.
Generate cryptographically random UUID v4 or time-ordered UUID v7, in bulk.
Encode and decode Base64 with correct UTF-8 handling, including the URL-safe alphabet.
Generate ULIDs — sortable by creation time like UUID v7, but Crockford Base32 instead of hex.