Modules
Shared voice infrastructure — STT/TTS behaviours, adapter implementations, and the VoiceBridge orchestrator.
Per-session orchestrator: inbound audio → STT → turn callback → TTS → outbound audio.
DynamicSupervisor for VoiceEngine.Bridge processes.
Runtime configuration for voice profiles.
HTTP client for ElevenLabs API. Shared by STT and TTS adapters.
ElevenLabs STT adapter. Buffers chunks locally, single API call on stream_end.
ElevenLabs TTS adapter.
Behaviour for speech-to-text adapters.
HTTP client for on-prem DGX Spark STT/TTS services.
DGX Spark Whisper STT adapter.
DGX Spark Orpheus TTS adapter.
Behaviour for text-to-speech adapters.