SS Toolbox
SS Toolbox
All data is processed locally

Image Color Extractor — Dominant Color, 8-Color Palette & HEX Swatches

Pick a poster, screenshot or photo and get its dominant color plus an eight-color palette, computed in the browser with median-cut quantization. Swatches show HEX and copy on click; there is no color-count control, no HSL or CMYK output, and no CSS variable export.

After selecting the image, the color will be automatically extracted, click the color to copy
#100100100
Main Color
Palette

What this color extractor outputs — and what it leaves out

  1. Dominant color plus an 8-color palette

    Every load runs median-cut quantization: the largest cluster becomes the dominant color, and the palette is always eight swatches. The image is shown at 16:9 with object-fit contain as a preview only.

  2. No color-count or format settings

    The palette size is hardcoded to 8 (the dominant color comes from a separate 5-color pass), and values are RGB internally, printed as lowercase HEX only. There is no HSL, CMYK, RGB text or CSS variable export.

  3. Click a swatch to copy the HEX

    Each circle writes its HEX value to the clipboard and shows a toast. The image itself is not clickable, so there is no eyedropper and no way to sample one specific pixel.

  4. Canvas-based, but local files only

    The chosen image is drawn onto a canvas and read with getImageData, entirely in the browser. There is no image URL field, so cross-origin canvas tainting never comes up — only the file you pick is analysed.

  5. Sampling skips near-white and transparent pixels

    Every 10th pixel is sampled, and pixels with alpha below 125 or near-white values (all three channels above 250) are ignored. White backgrounds and transparent areas therefore never dominate the result.

Image color extraction FAQ

How are the colors calculated?

With ColorThief's median-cut quantization (MMCQ). The dominant color is the first cluster of a 5-color pass; the palette is an 8-color pass. Quality defaults to 10, meaning every tenth pixel is sampled.

Can I change the number of colors?

No. Eight swatches are hardcoded, with no slider or input, and the dominant color comes from a separate 5-color quantization. The palette is replaced on every new file.

Which color formats can I get?

Only lowercase HEX, shown on each swatch and copied on click. There is no RGB, HSL or CMYK readout, no CSS custom properties, and no PNG or text export.

Can I use an eyedropper or pick a single pixel?

No. The image is a preview only; clicking a swatch copies its HEX. There is no pixel-level picking and no way to sample a point on the photo.

Is the image uploaded, and is it processed in canvas?

Everything runs in a local canvas and nothing is uploaded. The file is loaded through an object URL and displayed in a fixed 16:9 preview area, then drawn at natural size for analysis.

Will a cross-origin image taint the canvas?

The page only accepts local files; there is no URL input and getColorFromUrl is not used, so CORS tainting cannot happen. For a web image, download it first and upload the file.

Why are white and transparent areas ignored, and what about huge images?

Pixels with alpha below 125 and near-white pixels (all channels above 250) are skipped. Very large images are drawn at natural size and read in full, so they can be slow and memory-heavy; there is no resize or size limit.

Related Tools
SponsorMindElixir