Skip to main content

How It Works

TestivAI integrates seamlessly with your existing test framework through a simple function call. No complex dependencies required.

Simple Integration​

// In your test
testivaiWitness('homepage-test')

// That's it! TestivAI handles the rest

TestivAI captures visual snapshots and analyzes them for regressions, all without modifying your test logic.

Framework Support​

TestivAI works with all major test frameworks:

  • Cypress
  • Playwright
  • Selenium
  • WebdriverIO
  • Puppeteer
  • And more...

No framework lock-in - use what works for your team.

REVEAL Engine™​

TestivAI's proprietary REVEAL Engine uses intelligent multi-layer analysis to catch visual bugs while minimizing false positives:

Non-AI Analysis (Fast & Deterministic)​

  • Structural checks — detects missing or broken elements instantly
  • Visual comparison — catches style, layout, and positioning regressions
  • Perceptual analysis — filters out meaningless visual noise like anti-aliasing

AI-Assisted Analysis (Smart & Context-Aware)​

  • Intelligent change analysis — explains if changes are bugs or expected updates
  • Plain English reasoning — provides human-readable explanations for every flagged change

This multi-layer approach ensures only meaningful visual changes are flagged for review.

Chrome Setup​

Chrome must start with --remote-debugging-port=9222. The testivai init wizard adds this automatically for supported frameworks:

  • Cypress: Injects via cypress.config.js setupNodeEvents plugin
  • Selenium: Adds ChromeOptions arg in the generated helper file
  • WebdriverIO: Documents the goog:chromeOptions config
  • Puppeteer: Adds --remote-debugging-port=9222 to puppeteer.launch() args

Baseline Management​

First test run    → Creates NEW_BASELINE
Future runs       → Compared against baseline:
   No changes     → Test passes
   Changes found  → Analyzed for impact
   Needs review   → Flagged for your attention

The first time a test runs, it gets a NEW_BASELINE status (🆕) and becomes the reference point for future comparisons. This ensures accurate tracking from the very beginning.

→ Learn more about baselines and the approval flow → Learn about all test statuses