ColorsMine vs Colormind
If you are comparing these two, you want colours pulled out of a photograph or a reference.
The short version. Colormind is a generator that produces palettes from an image or from a learned model. ColorsMine starts where a palette ends: it turns colour into a token system and then measures whether that system is usable. They overlap less than the search suggests.
Use Colormind when
you only need the extraction and will do the system work yourself. That is a real job and this page is not going to pretend otherwise.
What ColorsMine adds
Deterministic extraction — the same image always gives the same palette — and everything downstream of it.
What ColorsMine actually produces
| From | You get |
|---|---|
| One brand hex | 8 primitive ramps and 57 semantic variables — surface, text, border, primary, secondary, accent, and success / info / warning / danger — in light and dark |
| An image | Ten semantic roles with an 11-step OKLCH scale each, deterministically: the same image always gives the same palette |
| An existing palette | A grade out of 100 across contrast, colour-blind safety, harmony, distinctness and tonal range, plus the single change that moves it most |
| A Figma collection | Every mode graded, every canvas layer scanned for real contrast failures, and a proof frame that renders in both themes |
| A repository | npx colorsmine check --min B --wcag — the same engine as a build gate |
What it is checked against
- WCAG 2.2 — 4.5:1 for body text (1.4.3), 3:1 for large text and non-text contrast (1.4.11), 7:1 for AAA.
- APCA Lc, which accounts for polarity and disagrees with the ratio often enough to be worth reading.
- Protanopia, deuteranopia, tritanopia, and greyscale.
Cost and access
ColorsMine is free, needs no account, and makes no network calls — the palette engine, the 10,368,048-palette catalogue and the grading all run locally. Nothing you open leaves your machine.
This page describes Colormind at the level of what kind of tool it is, not as a feature comparison — tools ship and a feature table written today is wrong by next quarter. Check its own site for what it does now.
Other comparisons
Try it on a colour you already have
Paste a hex and see the whole system it implies, graded, in about a second.
Open ColorsMine