[GitHub Trending] huggingface/speech-to-speech
Scored daily by a customisable AI persona to surface the most relevant engineering leadership news.
HuggingFace's open-source speech-to-speech for local voice agents, high novelty and strong community signal.
Hugging Face's speech-to-speech library provides a modular, low-latency voice-agent pipeline (VAD → STT → LLM → TTS) exposed via an OpenAI Realtime-compatible WebSocket API, with every component swappable. The LLM slot supports OpenAI-compatible protocols, allowing use of hosted providers, HF Inference Providers, or local servers like vLLM or llama.cpp for a fully open stack. Already powering thousands of Reachy Mini robots in production, it installs via `pip install speech-to-speech` and defaults to Parakeet TDT for STT, Qwen3-TTS for speech output, and an OpenAI-compatible LLM.
huggingface