Dear SenseVoice Team,

I'm developing a commercial voice typing application for Android using sherpa-onnx,
and I'm interested in using the SenseVoice model for multilingual speech recognition.

I've reviewed the license files and found:

• This repository's "View license" links to the FunASR repository
• FunASR states "MIT License" for the project, but also links to a separate
  "Model License" (FunASR Model Open Source License Agreement) for pretrained models
• It's unclear which license applies to the SenseVoice model weights

I found a similar question in FunASR#749 where commercial use was implicitly confirmed,
but I'd like explicit clarification for SenseVoice specifically.

Specific models in question (distributed via sherpa-onnx):

• sherpa-onnx-sense-voice-zh-en-ja-ko-yue-int8-2024-07-17
• sherpa-onnx-sense-voice-zh-en-ja-ko-yue-int8-2025-09-09
• sherpa-onnx-sense-voice-funasr-nano-int8-2025-12-17

Questions:

1. Which license applies to these SenseVoice model weights - MIT or the FunASR Model License?

2. Is commercial use permitted in a paid mobile application?

3. What attribution is required? (e.g., in an "About" or "Licenses" section of the app)

4. Section 4.2 of the Model License mentions "prohibited behavior" that can forfeit the license.
   Could you clarify what this means in practice?

Thank you for your excellent work on multilingual speech recognition!