A synthetic MediaPipe canonical face animates with facial action units, gaze, and blinks while cycling between example interaction profiles; five derived signal layers — prosody, speech timing, language, facial action units, and eye contact — update in sync. Hover layers to annotate the face.