e2e: Automate Web and Mobile App Testing with Natural Language Instructions

Describe test objectives in natural language, and an AI agent will manipulate the actual app to achieve them. This is a hybrid E2E testing framework that verifies results using locators and assertions.

Actions performed by the agent are recorded and replayed in subsequent runs without model calls. AI inference is only re-invoked when the app structure changes, significantly reducing test execution speed and costs.

It supports web engines based on Chromium, Firefox, and WebKit, as well as iOS and Android simulators. It internally leverages Playwright and agent-device, alleviating the burden of writing traditional scripts.

You can freely integrate API keys or local models to avoid dependencies. Although the API may change as it is pre-1.0 release, it is open to the open-source ecosystem under the Apache-2.0 license.