Image Color Extractor

Image color extraction analyzes the pixels in a photo, logo, or screenshot and pulls out the colors that best represent it as a palette — useful for identifying a logo's brand colors or picking a UI theme that matches a photo's mood. toco.tools's extractor runs entirely in the browser: pick how many colors you want, and it computes each color's HEX value and its share of the image. Under the hood it uses median-cut quantization, a deterministic algorithm. It starts with one box covering the full range of R, G, and B values in the image, then repeatedly finds the box whose widest color channel has the most spread and splits it in two at the midpoint of that range, continuing until the requested number of boxes exists. Each box's pixels are then averaged into a single representative color. Unlike k-means, which depends on random starting points, the same image and color count always produce the same result. The tool leads with area share rather than just a color list because proportions matter in practice: a color covering 40% of an image is likely the design's dominant background or brand color, while one covering only 2% is probably a decorative accent or noise. A practical workflow is to scan the bars from largest to smallest and treat the top entries as your brand color candidates. Note that JPEG compression artifacts or heavy gradients can sometimes split what should be one color into several close variants, so it's worth eyeballing the final palette before using it.