Free tool

Every colour a website uses, and the ones that nearly match.

Paste any address. The palette checker reads its stylesheets and lists every colour they declare, how often, and the near-duplicates: one colour typed twice, a shade apart.

Try · · ·

The stylesheet is only half of it.

The palette checker reads CSS as it’s written, so it counts colours a page may never show and can’t see styles set by scripts. EvenUI loads the page in a real browser, reads what every element actually renders, and finds off-system spacing, type, colour and components, each pinned to the page with its CSS file and line. It checks against your design system, and plugs into Claude Code.

Free for Mac and Windows. Pro audits whole sites: $34.99 this week with LAUNCH30, usually $49.99.Buy Pro

What counts as a near-duplicate colour?

Two colours less than 3 apart by CIEDE2000, the standard measure of how different two colours look. Below 1 most people can’t tell them apart at all; below 3 it takes two swatches side by side. On a real page they’re almost always one colour typed twice: #2463eb in one component, #2563eb everywhere else.

Why does it find more colours than the page shows?

It reads the stylesheets, and stylesheets carry colours for other pages, dark mode, hover states and whole token sets. EvenUI, the desktop app, counts only what renders, element by element. In our scan of 41 product homepages, the median page rendered 31 colours.

How do I fix near-duplicates?

Pick the one used most, replace the others with it, and better still, give it a name: a CSS variable or a design token. Copy the palette above with the duplicates merged for a starting point. To find exactly which elements use each stray colour, and the line of CSS behind it, scan the page with EvenUI.

What does it read, and what’s kept?

The page’s HTML, its linked stylesheets (up to 24, following @import), its style tags and inline styles. Nothing is stored except a cache of the result for an hour, so the same link loads fast for whoever you share it with.