| name | visual-regression-testing |
| description | Detect unintended visual changes in UI by comparing screenshots across versions. Use for visual regression, screenshot diff, Percy, Chromatic, UI testing, and visual validation.
|
Visual Regression Testing
Table of Contents
Overview
Visual regression testing captures screenshots of UI components and pages, then compares them across versions to detect unintended visual changes. This automated approach catches CSS bugs, layout issues, and design regressions that traditional functional tests miss.
When to Use
- Detecting CSS regression bugs
- Validating responsive design across viewports
- Testing across different browsers
- Verifying component visual consistency
- Catching layout shifts and overlaps
- Testing theme changes
- Validating design system components
- Reviewing visual changes in PRs
Quick Start
Minimal working example:
import { test, expect } from "@playwright/test";
test.describe("Homepage Visual Tests", () => {
test("homepage matches baseline", async ({ page }) => {
await page.goto("/");
await page.waitForLoadState("networkidle");
await expect(page).toHaveScreenshot("homepage-full.png", {
fullPage: true,
maxDiffPixels: 100,
});
});
test("responsive design - mobile", async ({ page }) => {
await page.setViewportSize({ width: 375, height: 667 });
await page.goto("/");
await expect(page).toHaveScreenshot("homepage-mobile.png");
});
test("responsive design - tablet", async ({ page }) => {
Reference Guides
Detailed implementations in the references/ directory:
Best Practices
✅ DO
- Hide or mock dynamic content (timestamps, ads)
- Test across multiple viewports
- Wait for animations and images to load
- Use consistent viewport sizes
- Disable animations during capture
- Test interactive states (hover, focus)
- Review diffs carefully before approving
- Store baselines in version control
❌ DON'T
- Test pages with constantly changing content
- Ignore small legitimate differences
- Skip responsive testing
- Forget to update baselines after design changes
- Test pages with random data
- Use overly strict thresholds (0% diff)
- Skip browser/device variations
- Commit unapproved diffs