AI Test Execution

Overview

AI execution runs the generated steps in a browser agent and records screenshots, logs, and video.

Execution flow

1. Launch browser and navigate to target URL.
2. Execute tool calls (navigate, click, type, assert, etc.).
3. Persist screenshots and video to the test run.
4. Emit progress events over WebSocket.

Stability guidance

• Use fast mode for flaky pages.
• Limit max steps to avoid infinite loops.
• Prefer deterministic assertions.

Results

• Execution status and duration.
• Screenshots and optional video.
• Error summaries and stack traces.