Color palette extractor
Pull the dominant colour and a palette of up to 16 colours out of any image, as HEX, RGB and HSL.
Or choose them. Files are processed in this tab and never uploaded to any server.
Match a slide deck to a brand's logo, build a theme from a photograph, or find out which blue a screenshot actually uses. The image is sampled in your browser and grouped into colours by median cut, the same algorithm color-thief uses, giving you the dominant colour and a palette ordered by how much of the image each covers. The output is ready to paste: HEX, RGB and HSL for every colour, a block of CSS custom properties, and a swatch image.
How it works
- The image is first scaled down to 256 pixels on its longest side, which averages out JPEG noise and the soft edges that would otherwise take palette slots of their own.
- Median cut keeps splitting the busiest group of colours along its widest channel; the last quarter of the splits favour groups that are both large and varied, so a small, vivid accent can still earn a slot.
- Transparent pixels are skipped, so a logo on a transparent background yields the logo's colours; near-white can be skipped too when a white backdrop dominates.
Where your data goes
Nowhere. This tool runs entirely in your browser: the text you paste is processed by the page and is never transmitted to a server or written to a log.
This tool is free and needs no account. Its results exist only in your open page and are not saved anywhere.
What it costs
This tool is free, with no sign-in and no points.
Common questions
- Why isn't the dominant colour the one I see most?
- Each palette colour is the average of a group of similar pixels, so a sky with a gradient becomes one mid-blue rather than its brightest shade, and a large but subtly textured background can outweigh a bold subject. Asking for more colours splits those groups more finely.
- Why does the palette differ slightly between browsers?
- Browsers decode and downscale images with slightly different filters and colour management, so the sampled pixels differ by a unit or two in some channels. The resulting colours shift by the same tiny amount — invisible, but enough to change a HEX digit.
- How many colours should I ask for?
- Five or six gives a usable brand palette. Ask for twelve to sixteen when you want to see the minor tones of a photo, and fewer when the image is a flat illustration with only a handful of real colours, where extra slots would just split one colour in two.
The open-source behind it
This tool is a self-contained implementation. lokesh/color-thief (MIT) does the same job as a library — if you need this behaviour inside your own program, start there rather than calling a web page.
lokesh/color-thiefAlso known as
- color palette from image
- extract colors from image
- dominant color finder
- image color picker
- palette generator from photo
- get hex colors from image