Stride AI launches modular voice engine targeting ~380ms latency and 0.2s native interruption for humanoid robots

The US team Stride AI has built a modular voice engine targeting ~380ms latency and supporting 0.2s native interruption, aimed at easing the "uncanny valley" feel in interactions between humanoid robots and real-time voice agents. The engine emphasizes flexible integration into different tech stacks and hardware environments. It is currently open for early pilots to technical teams, robotics developers, and developers, with a demo link for trial and feedback.