Skip to content
Quality Assurance

Ship with
Confidence

Semantic regression detection. Not pixel diffing. Nexver compares property values, generates tests, and tracks coverage at the design level.

Regression Detection

8 semantic categories

Every UI change is classified by what it means — not how it looks pixel-by-pixel.

Layout

Width, height, padding, margin, alignment, and positioning changes.

Color

Background, foreground, border, shadow, and gradient changes.

Typography

Font size, weight, family, line height, and letter spacing.

Visibility

Display, opacity, overflow, and z-index regressions.

Semantic

Purpose, role, label, and content meaning shifts.

Accessibility

ARIA roles, labels, contrast ratio, and focus order.

Spacing

Gap, margin-between, and container padding changes.

Interaction

Click targets, hover states, transitions, and animation.

Workflow

From diff to dashboard in seconds

01

Design Diff

Compare two versions of your product element by element using Nexver.

02

Classify

Each change is categorized into one of 8 regression types with severity scoring.

03

Generate Tests

Playwright tests are auto-generated from the diff using the 4-tier selector strategy.

04

Report

Results appear in the Catalyst QA dashboard with coverage maps and CI status.

Immune System

Test metadata embedded in designs

an elementQA and UIXTestHints embed test metadata directly inside canonical components. Every node knows how to test itself.

// UIXTestHints embedded in a canonical button
{
  "testHints": {
    "selector": "[data-testid='submit-btn']",
    "fallback": "[role='button'][aria-label='Submit']",
    "assertions": [
      "should be visible",
      "should be focusable",
      "should trigger form submission"
    ],
    "interactions": ["click", "keyboard:Enter"]
  }
}

Selectors

4-tier selector strategy

1st

data-testid

Explicit test identifiers — most stable, recommended for all interactive elements.

2nd

ARIA roles/labels

Accessibility attributes provide semantic selectors that survive refactors.

3rd

CSS selectors

Class-based selection when test IDs and ARIA are unavailable.

4th

XPath (fallback)

Structural path selectors — least stable, used only as last resort.

Zero regressions shipped

Catch UI bugs before they reach production. Design-level testing for every platform.