The colors that actually appear in your logo
Upload a logo or image and this page samples every pixel, groups near-identical colors, and returns the most common ones as CSS custom properties ready to paste into a stylesheet.
This is a frequency count, not a design judgment -- background color often wins by sheer pixel count, which is usually what you want for a brand palette anyway.
Related tools
For true k-means clustering (better on gradients), see the Dominant Color Extractor (K-Means).
Frequently Asked Questions (FAQ)
How does this extract colors from my logo?
Your image is downsampled to a 64x64 canvas. Every opaque pixel is rounded to a coarse color bucket, buckets are counted, and the most frequent buckets become your swatches -- each shown as the true average color of the pixels in that bucket.
Is my logo uploaded anywhere?
No. Sampling runs entirely in your browser via the Canvas API. The file never leaves your device.
Why does a small accent color sometimes get missed?
This is frequency-based: a color that covers a tiny fraction of the image can lose out to background colors and anti-aliased edge blends. If an accent color is missing, try a larger source image or fewer swatches to reduce dilution from near-duplicate buckets.
How is this different from the K-Means extractor?
This page counts frequency of quantized colors (a histogram). The Dominant Color Extractor (K-Means) clusters pixels by proximity and averages each cluster, which can merge similar shades more accurately on gradients.
