Agents that hold spoken conversations, combining speech-to-text, LLM reasoning, and text-to-speech so users can talk to them instead of typing. - View it on GitHub