Core Concepts
The Web Agent
The web agent is the browser executor that runs your tests, interpreting plain-language steps and driving a real browser.
The web agent is what runs your tests. For each step, it reads your plain-language intent, finds the right element on the live page, and performs the action in a real browser. Then it evaluates your assertions against what's on screen.
Because the agent resolves intent at run time rather than replaying fixed selectors, your tests adapt to UI changes and fail only when behavior genuinely breaks.
Every run is recorded as an execution, with screenshots and a written outcome.
You don't manage the agent directly. You write test cases and the agent runs them on demand.