MirrorMe Technology recently launched the VIVA dexterous hand and achieved the world’s first real-time human-robot four-hand piano duet, marking the VIVA times CADA milestone. The robot no longer merely plays accurately but plays well, moving from preset motions to a performing partner with artistic expression and creative ability.
On naming, MirrorMe says VIVA comes from the Latin root for life and echoes the musical term vivace, while CADA derives from cadenza, the virtuoso flourish. VIVA times CADA also nods to viva cada dia, Portuguese for enjoy each day.
VIVA cracks the impossible triangle of force, speed and dexterity. Power limits force against speed, and more degrees of freedom cap each actuator’s power. Benchmarking the human hand, MirrorMe used forward design: it collected pianist data, extracted design targets, then modelled and topologically optimised. Complex finger changes, multi-fingertip coordination and fine trajectory control now approach the human motion envelope.
The key is a tendon-plus-linkage composite transmission. Tendons put the actuator behind the finger for speed. linkages give stiffness and precise force. the composite reproduces the hand’s natural coordination so every key strike and lift follows a real player’s path. Joint speed exceeds 1,000 degrees per second and single-finger peak force reaches 33 newtons, covering the full volume range and hard techniques like the Can-Can.
What gives VIVA a mind is MirrorMe’s self-developed CADA music embodied model. If VIVA is the body, CADA is the brain: it closes the loop from music understanding and creation to embodied performance and feedback learning. A general model reads natural-language intent, a music-generation model writes melody and arrangement, a performance model turns expression into fingering, touch and timing, and VIVA plays it on a real piano.
Each performance is not an end. The system keeps sensing the actual sound, motion and result, compares target and reality, and adjusts, a loop of creation, performance, perception, evaluation and learning. The piece Viva la Vida: Nueva Vida performed at the launch was generated with the model and self-corrected by the robot. MirrorMe sees music as an ideal arena fusing cognition, creation, expression and precise action, where general intelligence becomes specialised skill and continuous feedback evolves it.
MirrorMe comes from a decade of robotics work at Zhejiang University and focuses on original coordination of body, embodied model and ecosystem. It built the world’s only whole-body coupled-dynamics design software, stretching performance boundaries, and its full-size humanoid platform extends the same principles beyond the hand.
Editor’s note: This is an adapted translation of the original OFweek report. It has been trimmed and restructured for readability for an international business audience. The full original (in Chinese) is at https://robot.ofweek.com/2026-09/ART-898890-8120-30704831.html.