Image Color Picker

Upload an image, hover for a live pixel magnifier, click to sample exact colors, and get an adjustable dominant palette with real pixel shares and accessibility data. Everything happens locally in your browser — your image is never uploaded anywhere.

Choose a different image

Extracted palette

Colors:

Shares are the percentage of sampled pixels closest to each color. W/B badges mark colors that can carry white or black text at WCAG AA.

Resolution

Color groups*

Avg brightness

Avg saturation

Dominant hue

Transparent

*Distinct 4-bit-per-channel color groups in the analysis sample — a quantized estimate, not an exact unique-color count.

Export palette

How the extraction works

The image is downscaled to a 100×100 analysis sample, every opaque pixel is grouped into a reduced color bucket (4 bits per channel — 4,096 possible groups), and the largest groups become the palette, each swatch averaged from the real pixels inside its group. The share percentages and the statistics come from the same pass. This is why different tools extract different palettes from the same photo: bucket size, sample resolution, and whether clustering happens in RGB or a perceptual space all change the result — there is no single "correct" palette, only different summaries. Transparent pixels (alpha below 50%) are excluded, so logos on transparent backgrounds work correctly.

A practical note for brand work: a photo's dominant colors are usually its backgrounds, not its subject. If you're extracting brand colors from a logo screenshot, click-sample the exact pixels with the magnifier instead of relying on the automatic palette — then verify text pairings in the contrast checker, because colors that look great in a photo routinely fail as UI colors.

Frequently asked questions

No. The image is decoded and processed entirely in your browser using the HTML5 Canvas API — it never leaves your device, which also means there are no file-size delays from uploading.

Anything your browser can display: PNG, JPG, WebP, GIF, SVG, and AVIF on modern browsers. Very large images are scaled to 1600px on the long edge before sampling, which keeps things fast without visibly affecting the extracted colors.

Exact — the magnifier shows an 11-pixel square around your cursor with the center pixel outlined, and clicking reads that precise pixel's value from the canvas. For images larger than 1600px the canvas is a scaled copy, so a sampled pixel represents the scaled image rather than the original file's exact pixel grid.

Yes — pixels that are more than half transparent are excluded from the palette and statistics, so a logo on a transparent background yields the logo's colors, not a phantom white or black. The stats row shows what share of the image was transparent.

Every extractor summarizes millions of pixels into a handful of swatches, and the summarizing choices differ: how colors are grouped, at what resolution, and in which color space. Our method is described in full above — predictable and fast. If a specific pixel matters, sample it directly instead.

Yes — tab to the image, press Enter to place the sample point at center, then move it one pixel at a time with the arrow keys. The sampled color is announced in a live region for screen readers.