Where Playwright meets intelligent agents.
AgentWright is an experimental JavaScript/TypeScript framework that brings agentic automation into the Playwright ecosystem.
It allows multiple AI-driven agents to planner, generator, and healer within the same browser session — perfect for exploring the future of autonomous testing.
This repository demonstrates how three different agents can collaborate in a Playwright-based testing environment.
| Agent | Purpose | Description |
|---|---|---|
| PlannerAgent | Exploration | planner explores the app and produces a Markdown test plan. |
| GeneratorAgent | Transformation | transforms the Markdown plan into the Playwright Test files |
| HealerAgent | Execution | executes the test suite and automatically repairs failing tests |
Each agent operates independently but shares a unified Playwright context, simulating a multi-agent test orchestration model.
# Clone the repo
git clone https://github.com/<your-username>/AgentWright.git
cd AgentWright
# Install dependencies
npm install
# Run sample test
npx playwright test