GitHub radar

Audio8 TTS: Compact Voice Cloning Model

Audio8_TTS is a 0.6-billion-parameter multilingual text-to-speech model with zero-shot voice cloning in 11 languages. It runs on a single CUDA GPU and is available on Hugging Face under Apache 2.0.

01Audio8-AI/Audio8_TTS 308Python

Audio8_TTS is a multilingual text-to-speech model built by the Audio8 team. It uses a DualAR architecture with 0.6 billion parameters to generate natural-sounding speech with zero-shot voice cloning — meaning it can mimic any voice from a short reference recording. Supports 11 languages including English, Chinese, Japanese, and European languages. Available on Hugging Face under Apache 2.0.

Why a vibe-coder should care

If you want your app or agent to speak in a specific voice — your own, a character's, or a brand voice — Audio8_TTS lets you do that from a single short recording, no large server required.

Open on GitHub