Hello OpenBMB/VoxCPM maintainers,
We are evaluating VoxCPM2 (voxcpm 2.0.3) for local TTS and fixed-reference voice cloning.
We understand that the VoxCPM2 code and model weights are released under Apache-2.0 and that the model is described as commercially usable. We would appreciate clarification specifically about the rights and obligations for audio generated by the model:
- Who owns audio generated by VoxCPM2?
- May users save and export generated audio?
- May generated audio be published or redistributed, including on social media or to third parties?
- May generated audio be used commercially?
- Do different conditions apply to fixed-reference Voice Clone or other speaker-conditioned output?
- What consent or authorization is required from the reference speaker?
- Is disclosure that the audio is AI-generated mandatory or recommended?
- Must Apache-2.0 attribution, model attribution, NOTICE content, or another statement accompany generated audio?
- Are there output-use restrictions beyond impersonation, fraud, and disinformation?
We are specifically trying to distinguish commercial permission to use the model from rights in the generated audio.
If possible, could an authorized maintainer provide clarification that we may retain as production-use documentation?
Thank you.
Hello OpenBMB/VoxCPM maintainers,
We are evaluating VoxCPM2 (voxcpm 2.0.3) for local TTS and fixed-reference voice cloning.
We understand that the VoxCPM2 code and model weights are released under Apache-2.0 and that the model is described as commercially usable. We would appreciate clarification specifically about the rights and obligations for audio generated by the model:
We are specifically trying to distinguish commercial permission to use the model from rights in the generated audio.
If possible, could an authorized maintainer provide clarification that we may retain as production-use documentation?
Thank you.