Demo implementation of AI-powered test automation for QA teams, built with the Claude AI SDK and Playwright.
What it does
Users supply a requirements document and a target app URL. The agent:
- Logs into the application autonomously
- Navigates and exercises the app’s key flows
- Generates verification reports in JSON, JUnit XML, or HTML formats
Key features
- Supports Chromium, Firefox, and WebKit
- CI/CD integration examples: GitHub Actions, GitLab CI, Jenkins, Azure DevOps
- Agentic test execution — no manual test scripts required
- Built on the Claude Agent SDK for intelligent navigation decisions
Tech stack
Python · Claude AI SDK · Playwright · MIT License