Commercial redistribution permission for UVR_MDXNET_9482.onnx

[rzarandona]

Hello @Anjok07 and @aufr33,

Thank you for developing and maintaining Ultimate Vocal Remover.

I am developing CueCart, a paid proprietary Windows audio application. We are evaluating the following model for optional offline vocal separation during audio ingestion:

Model: UVR_MDXNET_9482.onnx

Our intended use is:

• Run the model locally inside a background ingestion worker
• Separate vocals from music.
• Bundle or automatically download the model for use by paid CueCart customers
• Use the model commercially without exposing or selling the model separately
• Include the applicable MIT license notice and attribution to UVR and its developers

The UVR repository states that third-party application developers may use UVR models while honoring the MIT license and providing credit. For our production-compliance records, could you please confirm the following specifically for UVR_MDXNET_9482.onnx?

1. Was this exact model trained or released by the UVR development team?
2. Are the pretrained model weights covered by the MIT license statement in the UVR repository?
3. May the model be commercially used and redistributed with a paid, closed-source application?
4. May we distribute the model directly in our installer or download it from our server after installation?
5. Are there any additional attribution, copyright, training-data, or redistribution requirements?
6. Is the following attribution sufficient?

UVR_MDXNET_9482 model
Ultimate Vocal Remover project and developers
Used under the MIT License
MDX-Net architecture credits: Kuielab and Woosung Choi

For identification, the ONNX artifact we are evaluating is named:

UVR_MDXNET_9482.onnx

SHA-256:

9d78f8566fa8198065214ab628be1de966a500c57786695aa4b13e2b27a7727d

A brief written confirmation would help us complete our commercial-production model admission review.

Thank you very much.

Reinhard
CueCart