Hugging Face and Cerebras Ship an Open Real-Time Speech-to-Speech Pipeline on Gemma 4

Summary

Hugging Face and Cerebras demonstrated an open, modular real-time speech-to-speech pipeline built for natural, low-latency voice conversation. The pipeline chains Nvidia's Parakeet for speech recognition, Google DeepMind's Gemma 4 31B vision-language model for reasoning, Cerebras for accelerated inference, and Alibaba's Qwen3TTS for text-to-speech — each component inspectable and swappable by developers. The emphasis is latency: Cerebras' inference speed is positioned as what removes the multi-s