Extract Dominant Palette from Image

Drop any photo, UI screenshot, or moodboard to extract 3–12 dominant color swatches clustered perceptually in OKLab.

0 bytes sent

100% Client-Side — Verify 0 bytes sent in Network tab

  • 100% browser-local
  • Zero server transfers
  • CSS Color 4 precision
or drop a file here · Ctrl V to paste · or drop PNG / JPG / WebP / SVG locally — never sent to a server

Instant Before & After Specimen

OKLAB K-Means + Median Cut Quantizer
Input Samplecolor-target-grid.png
Synthetic 240x160 calibrated multi-swatch test image (PNG)
Expected OutputOUTPUT
Extracted 6 dominant swatches: #2563EB, #10B981, #F59E0B, #EF4444, #8B5CF6, #0F172A

How to use Image to Palette

Image to Palette — cleanly inside your local browser with zero server transfers.

  1. Select screenshot or photoOpen any local PNG, JPG, or WebP file directly inside HTML5 Canvas memory.
  2. Sample, extract, or simulateInspect pixels, tune OKLab cluster count, swap targeted hues, or test CVD vision modes.
  3. Copy, chain, or exportCopy ready CSS/JSON snippets, download binary swatch files, or pipe directly to related Colr tools.

Frequently asked questions

Why quantize pixels in OKLab instead of raw RGB?

Euclidean clustering in OKLab mirrors human perceptual distance, avoiding bloated duplicate shades in shadows.

Can I export the extracted swatches directly to Adobe .ase or CSS?

Yes — click Copy CSS, Tailwind @theme, or Download .ase / Procreate .swatches straight from results.