Vibe Testing: AI generates tests from screenshots, code & apps

What is Vibe Testing?

Vibe Testing is AI-powered exploratory testing that generates test scenarios from real artifacts instead of manual specifications. Instead of writing test cases by hand, you provide:

• 📸 Screenshots of your UI → AI analyzes buttons, forms, navigation
• 🌐 Web app URL → Playwright crawls pages, GPT-4o Vision reads UI
• 📦 Source code (ZIP) → AI parses functions, routes, API calls
• 📱 Android APK → AI extracts manifest, activities, permissions

Flasqo's AI understands context and generates test scenarios that match real user workflows.

Why Vibe Testing matters

Traditional testing requires manually writing test cases. Problems with this:

• Time-consuming: Writing comprehensive tests takes days
• Requires expertise: Testers need deep product knowledge
• Outdated quickly: UI changes → tests need rewriting
• Misses edge cases: Humans forget uncommon workflows

Vibe Testing solves this by:

• ✅ Instant generation: Tests ready in 30-60 seconds
• ✅ Zero product knowledge needed: AI infers from artifacts
• ✅ Adapts to changes: Re-upload screenshot → new tests
• ✅ Catches edge cases: AI considers scenarios humans miss

How Vibe Testing works

4 ways to use Vibe Testing in Flasqo

Method 1: Web app crawler (recommended for live apps)

Step 1: Go to Testing Types → Vibe Testing → Launch

Step 2: Select "Crawl Web App"

Step 3: Enter URL: https://yourapp.com

Step 4: Click "Start Crawl"
Flasqo uses Playwright to:
• Navigate through pages
• Click buttons and links
• Fill sample form data
• Capture screenshots of each page

Step 5: AI analysis (30-60 seconds)
GPT-4o Vision analyzes screenshots and generates:
• User flow scenarios (signup → login → dashboard)
• Form validation tests (empty fields, invalid emails)
• Navigation tests (broken links, 404 pages)
• Visual checks (missing images, overlapping text)

Step 6: Review & run generated tests
Flasqo shows test scenarios in plain English:
• "User signup with valid email should succeed"
• "Login with wrong password should show error"
• "Empty form submission should display validation"
Click "Run Tests" to execute

Method 2: Screenshot upload (best for design mockups)

Step 1: Select "Upload Screenshots"

Step 2: Drag & drop PNG/JPG files
• Upload multiple screenshots (login page, dashboard, checkout, etc.)
• Supports Figma exports, mobile screenshots, desktop screens

Step 3: AI analyzes UI elements
GPT-4o Vision identifies:
• Buttons ("Sign In", "Submit", "Cancel")
• Input fields (email, password, search)
• Links and navigation menus
• Error states and validation messages

Step 4: Generated test scenarios
Based on visual analysis, AI creates:
• Click tests for all buttons
• Form fill scenarios
• Navigation flow tests
• Accessibility checks (missing labels, low contrast)

Method 3: Source code upload (for API/backend testing)

Step 1: Select "Upload Source Code"

Step 2: Upload ZIP of your codebase
Supports: Node.js, Python, Java, Go, Ruby, PHP

Step 3: AI parses code structure
Flasqo analyzes:
• API routes (app.get('/users/:id'))
• Function signatures and parameters
• Database models and schemas
• Auth middleware and guards

Step 4: Auto-generated API tests
AI creates:
• Endpoint tests for all routes
• Parameter validation tests
• Auth/permission tests
• Error handling tests (404, 500)

Method 4: APK upload (Android app testing)

Step 1: Select "Upload APK"

Step 2: Upload .apk file

Step 3: AI extracts app structure
Flasqo decompiles and analyzes:
• AndroidManifest.xml (activities, permissions)
• Intent filters (deep links, URL schemes)
• Network endpoints (API URLs in code)
• Third-party SDKs (analytics, crash reporting)

Step 4: Generated mobile test scenarios
AI creates:
• Activity launch tests
• Permission request flows
• Deep link tests
• Network security config checks

Example: E-commerce app Vibe Testing

Input: Upload 3 screenshots (product page, cart, checkout)

AI-Generated Test Scenarios:

1. Product page:
   • Click "Add to Cart" → Verify cart count increases
   • Change quantity → Price updates correctly
   • Select out-of-stock variant → "Add to Cart" disabled
2. Cart page:
   • Click "Remove" → Item disappears from cart
   • Update quantity to 0 → Item removed
   • Apply invalid coupon → Error message shown
3. Checkout:
   • Submit with empty email → Validation error
   • Enter invalid card number → Payment fails gracefully
   • Complete purchase → Order confirmation displayed

All generated in 45 seconds from just 3 screenshots. No manual test writing.

What AI detects from screenshots

• Interactive elements: Buttons, links, dropdowns, checkboxes
• Form fields: Input types (text, email, password, number)
• Validation: Required fields, error messages, success states
• Navigation: Menu items, breadcrumbs, back buttons
• Visual states: Loading spinners, disabled buttons, tooltips
• Accessibility issues: Missing labels, low contrast text, unlabeled icons

Best practices

1. Upload key user flows

Don't upload random screenshots. Focus on critical workflows: signup, login, checkout, payment.

2. Include error states

Upload screenshots showing validation errors, network failures, empty states. AI learns what to test.

3. Combine with functional testing

Vibe Testing generates scenarios. Use Functional Testing to validate API responses and data correctness.

4. Re-run after UI changes

When you redesign a page, re-upload screenshots to regenerate tests. AI adapts to new UI instantly.

Vibe Testing vs traditional test writing

When to use Vibe Testing

Use Vibe Testing when:

• You have mockups/designs but no functional app yet
• You need tests quickly (demo tomorrow, launch next week)
• You're testing a third-party app (no source code access)
• Your UI changes frequently (Agile/continuous delivery)
• You want exploratory testing ideas

Use traditional testing when:

• You need precise API contract validation
• You're testing complex business logic
• Compliance requires specific test cases
• You have detailed test specifications already written