Track A (personal voice). Speech recognition and speech synthesis run in your browser; the bridge only sees text. Hold the button, speak, release.
Send