Windows GUI for music genre and style detection using locally deployed AI model.
Powered by MAEST project and their pre-trained model targeting 519 Discogs music styles.
Click on image to enlarge.
- .NET 10.0 Runtime - Windows x64
- Python 3.12 and Astral uv
- FFmpeg
- At least 3 Gb of available disk space (7 Gb via hardlinks)
- Download latest release and extract files anywhere.
- Launch
MusicClassification.GUI.exe. - On first launch: a settings window will show up, requiring to perform first setup.
- Optionally update "Runtime folder" path, otherwise 3 Gb files will end up in your %localappdata%.
- Press "Initialize..." and then "Save".
Music Classification Utility's application code and Python adapter are licensed under GNU Affero General Public License v3.0, see LICENSE.
Copyright © 2026 Sergey Syrtsev.
This project utilizes MAEST (Music Audio Efficient Spectrogram Transformer) by Pablo Alonso-Jiménez, Xavier Serra and Dmitry Bogdanov. First setup of this application downloads the upstream codebase (linked in runtime-manifest.json) into user's runtime\vendor directory. MAEST's source code is licensed under AGPL-3.0.
Analysis uses discogs-maest-30s-pw-129e-519l-swa.ckpt, which is downloaded during initialization from MAEST's release. The checkpoint is supplied unchanged and is not included in this repository or release.
Model weights retain their applicable upstream terms and are not re-licensed under this application's AGPL-3.0 license.