Skip to content

Clarification on commercial use and long-term reuse of VoxCPM2 Voice Design outputs #407

Description

@jcssii

Hello OpenBMB team,

I am evaluating VoxCPM2 for a local, open-weight TTS workflow using Voice Design (no reference audio / no voice cloning), and I would appreciate clarification on the downstream rights for generated audio and designed synthetic voice identities.

The README describes VoxCPM2 as "Fully Open-Source & Commercial-Ready" and states that the code and weights are released under Apache-2.0 and are free for commercial use. I would like to confirm how this applies specifically to Voice Design outputs.

Could you please clarify the following:

  1. Commercial use of generated audio

    • May audio generated locally with VoxCPM2 Voice Design be used in monetized or otherwise commercial public content, including YouTube, Douyin/TikTok, Xiaohongshu, advertising, sponsorships, affiliate/commerce content, and brand promotion?
  2. Long-term reuse of a designed synthetic voice identity

    • If a user creates a new synthetic voice only from a natural-language description, without intentionally imitating a real person and without using reference audio, may that designed voice be reused long-term as the consistent voice of a channel, product, or brand?
    • May the user retain and reuse the same description, seed, and generation settings to preserve that voice identity across future content?
  3. Attribution, disclosure, or notice requirements

    • Are there any additional attribution, AI-generated-content disclosure, notice, or redistribution requirements that apply specifically to Voice Design audio outputs?
  4. Scope: local open-weight inference

    • Do the answers above apply to local inference using the released open weights, and not only to an official demo, hosted service, or API?
  5. Third-party voice/personality rights

    • For a Voice Design output that is not intentionally modeled after, cloned from, or prompted to imitate a real person, does OpenBMB impose any additional voice-identity or personality-rights restrictions beyond applicable law and the Apache-2.0 license?

If these points are already covered by an official policy, FAQ, model card clause, or terms page, a link to the exact section would be very helpful.

Thank you for the clarification.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions