Color Theory LabExecutable color-system handbook

Research-informed color system

Generate color contracts that remain useful after the mood board.

Build independent light and dark systems in OKLCH. Resolve reference, semantic, and component tokens. Audit interaction states and data palettes. Keep only the profiles you choose.

01 · Generator

Define intent before selecting color.

The profile controls brand expression. The harmony controls relationships. The contrast target constrains the resulting semantic tokens.

Semantic palette

Click any swatch to copy its hex value.

Seed scale

Perceptually ordered lightness with controlled chroma.

Harmony candidates

Starting relationships, not automatic semantic assignments.

02 · Playground

Inspect the system in context.

A palette is not a row of swatches. This playground applies the active profile to navigation, cards, forms, states, charts, and hierarchy.

Light profile

Portfolio overview

Signal first. Decoration remains subordinate.

Active users

Last 30 days

48.2K
↑ 12.4% · Healthy

Automation coverage

Core workflows

78%
↑ 4.8% · On track

Engagement trend

Hue separates series. Lightness communicates magnitude.

Form controls

Surface, border, text, and focus must remain distinct.

Semantic states

Text labels prevent color-only meaning.

✓ Healthy · All systems operational
! Warning · Capacity above 80%
× Error · Deployment failed
i Info · Maintenance Saturday

03 · Component states

Validate the state machine, not only the resting screenshot.

The same component token contract is exercised across default, hover, pressed, focus-visible, disabled, invalid, read-only, checked, and semantic-message states.

Interactive state matrix

Simulated states are labeled. Native disabled, read-only, invalid, and checked semantics remain real.

04 · Data palettes

Match palette family to data semantics.

Sequential scales encode magnitude. Diverging scales encode movement around a meaningful midpoint. Qualitative scales separate unordered categories.

Sequential

A single hue family with ordered lightness.

Diverging

Two directions around a neutral midpoint.

Qualitative

Distinct categories with redundant labels.

Semantic states

Familiar associations, independently contrast-adjusted.

Palette diagnostics

Perceptual-distance checks are diagnostic thresholds, not WCAG conformance criteria.


DiagnosticObservedThresholdResult

Color-vision simulations use fixed matrix approximations at full severity. They expose likely collisions but do not certify usability or replace testing with representative users.

05 · Profiles

Store only what you choose.

Named profiles and the optional remembered workspace stay in this browser. No network service is used.

Preference storage

Explicit local persistence with portable exports.

0 of 30 local profile slots used.

Current versus saved profile

Compare resolved token values before loading or replacing a profile.

Choose Compare on a saved profile.

06 · Tokens

Move from color values to system contracts.

Export semantic CSS custom properties or a structured JSON profile. Components should consume roles, not raw palette steps.

Token dependency architecture

Values resolve inward. Components never consume palette values directly.

Portable token contract

Export implementation CSS, a resolved profile, a DTCG 2025.10 token set, a multi-context resolver, or typed application constants.

07 · Principles

The theory behind the controls.

These rules convert color research into implementation constraints.

01

Separate the channels

Recommended

Hue distinguishes categories. Lightness carries hierarchy and order. Chroma controls emphasis.

02

Generate modes independently

Normative

Light and dark appearances need separate surfaces, borders, text, chroma, and elevation decisions.

03

Semantics before harmony

Normative

A harmonious palette can still fail contrast, focus, states, forced colors, and category discrimination.

04

Avoid false boundaries

Recommended

Ordered data needs perceptually ordered scales. Rainbow scales can distort magnitude and create artificial transitions.

05

Never encode with color alone

Normative

Pair color with text, iconography, shape, position, pattern, or numeric labels.

06

Audit in context

Normative

Contrast ratios are a floor. Review grayscale, common color-vision deficiencies, real devices, and task performance.

08 · Handbook projection

Trace guidance to implementation and tests.

The SPA is an executable projection. The handbook remains canonical. Every exported profile carries the guidance IDs that constrained generation and validation.

Guidance coverage registry

Status separates normative requirements from recommendations and experimental diagnostics.

Guidance IDStatusImplementationVerification

09 · Evidence

Primary and implementation sources.

The application is derived from the Visual Design Research System v1.3 and current standards guidance.

ColorBrewer

Sequential, diverging, and qualitative palette families for mapped and quantitative data.

Viridis color maps

Implementation precedent for perceptually ordered, grayscale-readable continuous scales.

Save color profile

This profile will be stored in local browser storage. The app never silently discards profiles when the 30-profile limit is reached.