Voice

Real-time speech-to-text in the chat composer. The user speaks, the runtime transcribes, the agent runs the resulting prompt.