Chat Interface

Create tests by describing them in plain English.

Chat Interface

The chat interface is where you describe what you want to test, and AI generates the test for you. No coding required.

How It Works

  1. Open a project that has been explored
  2. Click Chat to open the chat interface
  3. Describe what you want to test in plain English
  4. The AI generates a Playwright test based on your request

Writing Effective Prompts

Be Specific

❌ "Test the homepage"
✅ "Verify the homepage shows the hero section with a 'Get Started' button"

Reference Real Elements

The AI knows your app from exploration. Reference actual elements:

✅ "Click the 'Sign In' button in the header"
✅ "Fill the email field with 'test@example.com'"
✅ "Verify the 'Welcome back' message appears after login"

Describe the Flow

For multi-step tests, describe the complete flow:

✅ "Test the checkout flow:
   1. Add a product to the cart
   2. Go to checkout
   3. Fill in shipping details
   4. Verify the order confirmation page"

Example Prompts

Login Flow

Test the login flow with valid credentials.
Enter 'user@example.com' as email and verify we reach the dashboard.

Form Validation

Test that the signup form shows an error when:
- Email is empty
- Password is too short

Navigation

Verify that clicking 'Products' in the navigation
takes us to a page with a product grid.

E-commerce

Test adding a product to the cart:
1. Go to the product page
2. Click 'Add to Cart'
3. Verify the cart badge shows 1 item

Understanding the Response

When you submit a prompt, the AI:

  1. Analyzes your request against the explored flow graph
  2. Identifies relevant screens and elements
  3. Generates Playwright test code
  4. Explains what the test does

You'll see:

  • The generated test code
  • A description of each step
  • Any assumptions the AI made

Iterating on Tests

You can refine tests through conversation:

User: "Test the login flow"
AI: [generates test]

User: "Also verify the 'Remember me' checkbox works"
AI: [updates test to include checkbox verification]

User: "Add a test for invalid password"
AI: [adds negative test case]

Running Generated Tests

After generating a test:

  1. Click Run Test to execute immediately
  2. View results including screenshots and timing
  3. If it fails, iterate with the AI to fix it

Promoting Tests

When you're happy with a test:

  1. Click Promote to save it
  2. Promoted tests run automatically (schedule or CI/CD)
  3. You can manage promoted tests in the Tests tab

Tips for Better Results

  1. Reference the flow graph: Check what screens/elements exist
  2. Start simple: Test one thing at a time, then combine
  3. Be explicit about verification: Tell the AI what success looks like
  4. Iterate: Use conversation to refine the test

Next Steps