Action Commands

Action Commands represent the core interaction layer in Drizz. They are used to perform user-driven actions such as tapping, typing, scrolling, dragging, and validating content on the screen. Drizz interprets these commands using Vision AI and context-aware reasoning to reliably identify UI components without relying on fragile selectors or DOM dependencies.

These commands are authored in natural language and mapped to deterministic actions during execution. Each Action Command corresponds to a single user intent, ensuring predictable behavior across Android, iOS, and mobile web applications. By combining visual understanding, contextual cues, and execution heuristics, Drizz delivers highly stable UI interactions even in dynamic, image-heavy or frequently updated interfaces.