Higgs Audio
A local text-to-speech model creates expressive narration and multilingual dialogue, with voice cloning and native Hugging Face Transformers support.
Higgs Audio's Higgs TTS 2 is a downloadable speech model for developers building narration and conversational audio into applications. It adapts spoken delivery to the text and can generate conversations with distinct voices.
- Expressive narration: Generate speech with pacing and intonation that respond to the text's phrasing and emotional tone.
- Multilingual dialogue: Create conversations across multiple languages, with written speaker descriptions guiding the voices.
- Voice cloning: Use reference recordings to reproduce a speaker's voice without training a separate model, including dialogue with two cloned voices.
- Music and humming: Generate speech together with background music, or melodic humming in a cloned voice.
Higgs TTS 2 runs locally through Hugging Face Transformers. It's a model developers can integrate into their own software. Higgs Audio V2 and Higgs TTS 2 are names for the same model.
The repository code uses Apache 2.0, while the model weights have separate license terms. Boson AI also offers a hosted API for the separate Higgs Audio v3 model; that service doesn't run on your hardware.
Logotyper
