SmartTest Studio

AI-powered Playwright test generator that analyzes web pages and produces executable test code.

What It Does

Provide a URL, get back working Playwright tests. The system:

1. Crawls the page with Playwright to extract forms, buttons, and inputs
2. Sends page structure to GPT-4 to generate test scenarios
3. Converts scenarios into executable Playwright TypeScript code
4. Runs tests in Chromium and captures screenshots
5. Returns results with generated code

Tech Stack

• Frontend: React, TypeScript
• Backend: Node.js, Express, TypeScript
• Testing: Playwright
• AI: OpenAI GPT-4

Quick Start

# Clone and install
git clone https://github.com/yourusername/smarttest-studio.git
cd smarttest-studio

# Backend
cd backend
npm install
npx playwright install chromium
cp .env.example .env
# Add OPENAI_API_KEY to .env

# Frontend
cd ../frontend
npm install

# Run (two terminals)
cd backend && npm run dev
cd frontend && npm start

Visit http://localhost:3000

Project Structure

smarttest-studio/
├── backend/
│   ├── src/
│   │   ├── routes/api.ts           # Express API
│   │   ├── services/
│   │   │   ├── crawler.ts          # Page analysis
│   │   │   ├── aiGenerator.ts      # GPT-4 integration
│   │   │   ├── codeGenerator.ts    # Code generation
│   │   │   └── testExecutor.ts     # Test execution
│   │   └── types/
│   └── package.json
└── frontend/
    ├── src/
    │   ├── components/              # React UI
    │   └── api/client.ts
    └── package.json

API

POST /api/test-generation-flow

// Request
{ "url": "https://example.com/login" }

// Response
{
  "pageAnalysis": { ... },
  "testScenarios": [ ... ],
  "playwrightCode": "import { test } from '@playwright/test'...",
  "results": { ... },
  "screenshots": [ ... ]
}

Example Output

Input: https://the-internet.herokuapp.com/login

Generated code:

import { test, expect } from '@playwright/test';

test('Valid login credentials', async ({ page }) => {
  await page.goto('https://the-internet.herokuapp.com/login');
  await page.fill('#username', 'tomsmith');
  await page.fill('#password', 'SuperSecretPassword!');
  await page.click('button[type="submit"]');
  await expect(page.locator('.flash.success')).toBeVisible();
});

Limitations

MVP scope:

• Single-page analysis
• No auth/persistence
• Local execution only
• Basic error handling

License

MIT

Note

Built for Stably AI application to demonstrate Playwright and LLM integration.