Full-duplex speech models, frontend-backend architectures, tool calls, and voice-agent tasks.
Demo for Duplex S2S Frontend-Backend model