Run open-source TTS on your Amazon EKS cluster using the OmniVoice (600+ languages) model with KServe and the vLLM-Omni serving engine, streaming PCM over WebSocket below 0.1 RTF. - View it on GitHub