Installation

pnpm install
pnpm build

Usage

Exploring Scenarios

Explore scenarios and generate tests using LLM and MCP.

pnpm start:explore -- [options]

Options:

• --scenario, -s: Scenario to run (Required)
• --url, -u: Base URL to start from (Required)
• --api-key, -k: LLM API key (Gemini by default, Required)
• --input, -i: User input (key,value,description format)
• --domain-context, -d: Domain context
• --max-attempts, -m: Maximum number of attempts
• --precondition, -p: Precondition file name to run before scenario(e.g., "example" for preconditions/example.spec.js)
• --expectation, -e: Expected outcome or result of the test scenario
• --steps: Detailed test steps (format: string array)

Example:

pnpm start:explore --scenario "User can access dashboard page after login" \
  --url https://www.google.com \
  --api-key API_KEY \
  --input email,test@email.com,"account email" password,test123,"account password" \
  --expectation "User should be able to see the dashboard with welcome message" \
  --steps "Go to login page" "Enter credentials and click login" "Verify dashboard is displayed"

Using YAML Configuration Files

You can also use YAML configuration files to define more complex test scenarios with detailed expectations and steps:

apiKey: "your-api-key"
maxAttempts: 50
baseUrl: "https://example.com"
inputs:
  - key: "email"
    value: "test@example.com"
    description: "User email for login"
domainContext:
  - "This is an e-commerce website"
browserOptions:
  browser: "chromium"
  headless: false
testContexts:
  - scenario: "User login and dashboard access"
    expectation: "User should successfully login and see personalized dashboard"
    steps:
      - "Navigate to login page"
      - "Enter login credentials"
      - "Click login button"
      - "Verify dashboard content is displayed"

Run with YAML file:

pnpm start:explore-file --file path/to/config.yaml

Running Recorder

Use the Recorder to capture user actions on web pages and generate test code. It is Intended for creating preconditions for test scenarios. (Simple wrapper around the Playwright Test Generator)

pnpm start:record --url <URL> [options]

Options:

• --url, -u: URL of the web page to record (required)
• --output-file, -o: Output file path (optional, default: ./preconditions/recording-{timestamp}.spec.js)
• --language, -l: Output language (optional, default: playwright-test)
  • Supported languages: playwright-test, javascript, python, java
• --headless: Run in headless mode (optional, default: false)

Examples:

# Record with default settings
pnpm start:record --url https://www.google.com

# Specify output file
pnpm start:record --url https://www.google.com --output-file ./my-test.spec.js

# Output in Python
pnpm start:record --url https://www.google.com --language python

# Run in headless mode
pnpm start:record --url https://www.google.com --headless