Hlas-1

Hlas-1 is an advanced open-source text-to-speech model that supports expressive and multilingual speech synthesis including emotions, accents, and vocalizations. It features zero-shot voice cloning from short reference clips, runs on a single 12 GB GPU, and provides studio-quality output in multiple languages. The model is based on Breeze TTS 2 architecture and allows detailed performance tagging inline with the text input.