Releases: NeptuneHub/AudioMuse-AI
AudioMuse AI v0.8.12: ARM build fix
Release Date: February 22, 2026
AudioMuse-AI v0.8.12 fix the arm build image as issue #343
What's Changed
- Devel -> Main - fix ARM build workflow by @NeptuneHub in #344
- Implement concurrency and update image tag handling by @NeptuneHub in #345
Full Changelog: v0.8.11...v0.8.12
AudioMuse AI v0.8.11: Minor improvement
Release Date: February 22, 2026
AudioMuse-AI v0.8.11 introduce minor improvement to the GMM artist index update at the end of the analysis.
Full Changelog: v0.8.10...v0.8.11
AudioMuse AI v0.8.10: Minor improvement
Release Date: February 15, 2026
AudioMuse-AI v0.8.10 includes minor improvements in voyager Song and Artist index to avoid bug on big music collection (> 150k songs).
This release also include an improvement on Navidrome (open subsonic API) to correctly share the client name.
AudioMuse AI v0.8.9: Minor improvement
Release Date: February 11, 2026
AudioMuse-AI v0.8.9 includes minor improvements to batch task cancellation.
Batch operations (Analysis, Clustering, and Cleaning) now enforce cancellation of all running tasks and reconcile the task database to ensure a consistent state.
AudioMuse AI v0.8.8: Dark modes
Release Date: February 01, 2026
AudioMuse v0.8.8 introduces Dark Mode, along with several performance and usability improvements. Artist index rebuilding is now smarter, updating only artists affected by analysis or cleaning processes. Album names have been added, and multiple other minor fixes and refinements are included to enhance overall stability and experience.
An additional experimental -noavx2 image will also be released for users with older CPUs that do not support AVX2. This version relies on legacy dependencies to maintain compatibility and is not the recommended setup, but is provided as a best-effort solution to keep AudioMuse accessible on older hardware. On this you can read more here: #305
What's Changed
- Devel -> Main - Added Album name by @NeptuneHub in #291
- Revise GPU deployment documentation for clarity by @NeptuneHub in #303
- Fix: update artist_gmm_manager by @NeptuneHub in #304
- Feature/dark mode by @rendyhd in #292
Full Changelog: v0.8.7...v0.8.8
AudioMuse AI v0.8.7: Dependencies update and improvement
Release Date: January 19, 2026
AudioMuse v0.8.7 include minor bugxix and update but also major dependencies update. This new release aim to improve the maintainability of the code in the future.
What's Changed
- Devel -> Main - Clap student update by @NeptuneHub in #278
- Optimize Docker image: Ubuntu 24.04, smart model caching, remove unused PyTorch dependencies by @Copilot in #269
- Fix file extensions not being detected for audio from Lyrion Media Server by @AlexCSDev in #282
- Dependencies update by @NeptuneHub in #281
- Devel -> Main - dependencies update and minor improvement by @NeptuneHub in #288
New Contributors
- @AlexCSDev made their first contribution in #282
Full Changelog: v0.8.6...v0.8.7
AudioMuse AI v0.8.6: Minor bug fix
Release Date: January 4, 2026
AudioMuse v0.8.6 include minor bugfix to song alchemy and AI clustering naming functionality.
What's Changed
- add option to change default port with docker compose deployement by @NuxitBrain in #255
- Fix artist IDs being added to playlists in Song Alchemy by @Copilot in #258
- Add parameter fallback logic for APIs rejecting temperature/max_tokens by @Copilot in #260
- Devel -> main by @NeptuneHub in #268
New Contributors
- @NuxitBrain made their first contribution in #255
Full Changelog: v0.8.5...v0.8.6
AudioMuse AI v0.8.5: analysis improvement
Release Date: December 27, 2025
AudioMuse v0.8.5 continues the work started in #225, with further improvements to the stability of CLAP analysis.
We introduced a new environment variable, PER_SONG_MODEL_RELOAD, which is enabled by default. When active, both the Musicnn and CLAP models are reloaded before each song inference.
The main benefit of this approach is improved memory stability: it prevents RAM usage from continuously increasing, at the cost of slightly slower analysis. We believe prioritizing stability over speed is the right choice.
A big thank you to everyone who helped test this release and shared their results, your feedback is invaluable in making AudioMuse-AI better every day.
What's Changed
- Add standalone student CLAP distillation framework by @Copilot in #252
- Devel -> Main - CLAP mini batch by @NeptuneHub in #253
Full Changelog: v0.8.4...v0.8.5
AudioMuse AI v0.8.4: CLAP analysis stability improvement
Release Date: December 25, 2025
AudioMuse v0.8.4 continue follow up on #225 keeping improving the stability of the CLAP analysis. This is particulary usefull for the user of GPU with -nvidia image that have OOM issue.
We also worked on readme.md by keeping the most important information needed for new user and by moving and improving the detailed one in /docs. Of particulary interest is /docs/ARCHITECTURE.md that explain how the different component of AudioMuse-AI work together.
What's Changed
- Fix GPU OOM errors with improved memory allocation and CPU fallback by @Copilot in #242
- Fix PostgreSQL NUL byte errors and ONNX GPU memory allocation failures by @Copilot in #245
- Fix memory leaks on non-OOM errors and enhance string sanitization by @Copilot in #246
- Fix OOM fallback to CPU and memory cleanup in ONNX inference by @Copilot in #248
- [WIP] Fix errors in CI logs for memory cleanup tests by @Copilot in #249
- Devel —> Main by @NeptuneHub in #247
Full Changelog: v0.8.3...v0.8.4
AudioMuse AI v0.8.3: Text Search resource usage improvement
Release Date: December 22, 2025
AudioMuse v0.8.3 follow up on #225 in regard of GPU memory usage during the Text Search functionality. This will improve the memory usage and the speed of the analysis on both GPU and RAM.
We also improved the deployment.yaml and the docker-compose.yaml example to show the usage of the env var CLAP_ENABLED. For slow cpu it can be put as false to disable the text search functionality and skip the CLAP analysis.
What's Changed
- Configure NVIDIA GPU resources in docker-compose by @NeptuneHub in #228
- Remove unused imports and environment variables by @Copilot in #229
- Fix duplicate CLAP_PYTHON_MULTITHREADS parameter in README by @NeptuneHub in #236
- Mulan -> devel by @NeptuneHub in #240
- Devel -> Main - CLAP fix and improvement by @NeptuneHub in #241
New Contributors
- @Copilot made their first contribution in #229
Full Changelog: v0.8.2...v0.8.3