Our voice carries our health condition long before symptoms appear. I will build a soft-robotic larynx that mimics voice. A high-speed camera, sensors, and motorized stages will control and measure the model, which will reproduce healthy and diseased phonation. Data will clarify how tissue motion creates sound and support early-warning biomarkers, e.g., for Parkinson’s. I hypothesize that the soft-robotic larynx will predict vocal qualities and early pathology which are not measurable in humans.