Voice recognition is handled by your browser/device; audio may be processed by its platform speech service.