Serverless TTS/STT at the edge.
Max speaks and listens through Workers AI — MeloTTS for speech and Whisper for transcription — so voice works without any local dependency.