Pinned post
Speech-to-speech no longer needs speech-to-text!
Until now, our stack was VAD -> STT -> LLM -> TTS. Now it can send audio directly to multimodal LLMs:
VAD → MLLM → TTS
No STT. The model understands your voice.
Now go build better voice agents!
github.com/huggingface/sp…
Until now, our stack was VAD -> STT -> LLM -> TTS. Now it can send audio directly to multimodal LLMs:
VAD → MLLM → TTS
No STT. The model understands your voice.
Now go build better voice agents!
github.com/huggingface/sp…
1:18