Image Palette Extractor
Tip: colours are clustered with the median-cut algorithm. A larger sampling step is faster but slightly less precise. Click any swatch to copy its value.
How to Use
📖 About this tool
Upload any image and the tool clusters pixel colours with the median-cut algorithm, extracting 2–24 dominant colours sorted by share of total pixels. Each colour is shown as HEX, RGB and HSL; click a swatch to copy. Copy all values at once or export a CSS custom-property file for your front-end project. The sampling step trades speed against precision. Everything runs locally. Everything you drop into this tool stays on your own device. The whole job is done by your browser using standard web technologies, so no file is ever uploaded to a server, no account is required, and nothing is stored after you close the tab. That matters when the material is private: screenshots that contain personal information, unreleased product images, recordings of internal meetings, or client footage you are not allowed to share. It also means the tool keeps working offline once the page has loaded, which is useful on a slow connection or on a machine behind a strict firewall. Because processing happens on your own hardware, speed depends on your device: a modern laptop handles large files quickly, while an older phone may need noticeably longer. For the best results, work with the highest quality source you have, since every conversion step can only preserve detail that was already there. If an output looks wrong, try reloading the page and running it again before assuming the file is broken. The interface adapts to the width of the panel it is placed in, so it works the same way on a desktop screen, a tablet and a phone. All controls are labelled in the language of the site you are visiting, and the tool never asks for permissions it does not need. Typical uses include building a colour scheme from a reference photo, generating theme variables for a website, picking brand colours for a poster, and comparing the tone of several images. Results are ordered by how much of the image each colour covers.
✨ Features
- Median-cut clustering：A classic quantisation algorithm — far more accurate than plain averaging
- Sorted by share：Colours ranked by how much of the image they cover
- Three colour formats：Switch between HEX / RGB / HSL; click a swatch to copy
- CSS variable export：Export CSS custom properties ready for front-end use
- Tunable sampling：A larger step is faster; lower it when you need precision
🚀 How to use
- 1. Upload an image;
- 2. Set colour count and format;
- 3. Click Extract;
- 4. Click a swatch to copy or download CSS variables.