name: testing-visual-regression
description: Visual regression (baseline diffs, UI stabilization)
List of pages/components for visual coverage
Baseline version and update rules
Browsers and viewports
Dynamic areas to mask
Diff sensitivity threshold
Theme/locale requirements
Stabilize data and state (fixtures/seed)
Disable animations and auto-updates when needed
Fix fonts, theme, and locale
Fix viewport size and scale
Screenshots are reproducible and deterministic
Comparison is done against the correct baseline
Dynamic areas are masked or stabilized
Changes are confirmed by a human (review)
The tool and comparison version are recorded
- Critical screens and user flows
- High-variance UI components
- Tables, forms, charts, modals
- Different viewports and themes
Expected output and baseline match by version
Diff threshold is documented
Results include artifact links
Reasons for baseline updates are recorded
- Do not update baseline without reviewing changes
- Do not compare screenshots with unstable data
- Do not mix different locales/themes in one baseline
Compare a product card across 3 viewports with fixed data
Verify visual changes in a modal after an update
By OzeroHAX schedule Updated 2/16/2026