FireRedTeam

/

FireRedTTS3

Like 76

Follow

FireRed Team 835

Hi, thank you for releasing FireRedTTS3.

GitHub and Hugging Face list Apache-2.0, but the Usage Disclaimer says zero-shot voice cloning is intended solely for academic research.

Could you clarify whether this is a binding usage restriction or a statement of intended use?

Specifically, is commercial use permitted for:

1. Speech generated through voice cloning, using reference recordings we have permission to use?
2. Original voices generated through FireRedTTS3-Instruct voice design?
3. Using those voice-design outputs as reference audio for FireRedTTS3-Base?

Our intended use is monetized livestreams and video clips. We would run the models locally without distributing their weights.

Does Apache-2.0 cover both Base and Instruct weights, and are there additional terms governing generated audio?

Thank you for clarifying.