- require exact model matches and preserve links for unavailable custom models
- restore complete stems from model configs and validate downloaded assets
- align ensemble channels and batches without dropping audio data
- stream uploads safely and restore localized model names
- add Python and frontend regression coverage
* feat: delegate custom model architecture detection to pymss
Default custom separation nodes to model_type=auto while preserving saved manual selections. Delegate catalog detection and runtime selection to the shared upstream implementation. Keep detection errors in ComfyUI's standard execution error flow and cover config preservation, explicit overrides and real inference.
Requires publication of pymss 2.1.6 and pymss-core 0.1.9 before standard dependency installation is available. Validation: 8 node tests passed, plus full PolarFormer weights through CPU and CUDA node execution.
* fix: expose conformer architectures in custom model nodes
* Update README
Removed BS PolarFormer section and clarified existing workflows.
---------
Co-authored-by: Sucial <1584846096@qq.com>
- Removed the unsafe `PymssLoadAudioBatch` node and its related registrations, locale entries, and example workflow.
- Kept the `Upload Audio` button, but restricted `/comfy-mss/upload-audio` to local clients only (`comfy_mss/services/routes.py:30-36`).
- Uploads are always saved under ComfyUI's input directory, with filename validation and `realpath`/`commonpath` checks to prevent path traversal.
- Migrated node and button translations to ComfyUI's official locale system using `locales/<language>/main.json` and `/i18n`.
- Add compatibility for ModelEntry without config_instruments
- Fix Windows DirEntry sorting when scanning custom model folders
- Update custom model discovery to use folder-based MSST model layout
- Require a model weight file and YAML config in each custom model folder
- Exclude VR/UVR models from Custom MSS Separate
- Update custom model path and layout documentation
- Introduced `paths.py` for handling model directory resolution and registration.
- Created `catalog.py` to manage model metadata and stem extraction.
- Implemented `routes.py` for handling audio uploads and model retrieval via HTTP.
- Added utility functions in `audio.py` for audio conversion, loading, and saving.
- Developed ensemble functions in `ensemble.py` for combining audio inputs.
- Enhanced web interface with new JavaScript files for audio loading, separation, and ensemble node management.
- Defined constants for audio types and node configurations in `constants.js`.
- Improved user experience with dynamic input/output management in audio processing nodes.
Sucial