Skip to content

Remove legacy Python WebUI - #229

Merged
0xShug0 merged 1 commit into
0xShug0:mainfrom
mirek190:agent/remove-legacy-python-webui
Aug 13, 2026
Merged

Remove legacy Python WebUI#229
0xShug0 merged 1 commit into
0xShug0:mainfrom
mirek190:agent/remove-legacy-python-webui

Conversation

@mirek190

Copy link
Copy Markdown
Contributor

Summary

Make the embedded native Svelte UI the repository's only browser interface and remove the superseded Python/Gradio and legacy realtime WebUI implementations.

Changes

  • remove the Python/Gradio application, launch scripts, environment helpers, Python dependencies, tests, and generated-output placeholders
  • remove the separate legacy realtime Python server and static browser client
  • remove duplicate legacy voice assets now supplied by the native UI's embedded demo-voice bundle
  • simplify webui/README.md around audiocpp_server --ui
  • update root documentation, the SenseVoice contributor notes, the voice-directory example, packaging, ignore rules, and catalog comments
  • regenerate the checked-in native single-file UI artifact from the current source

Scope boundary

This deliberately keeps tools/model_manager_v2.py and tools/model_manager_deprecated.py. They are repository-wide model preparation utilities used by the native server's Models workflow, not browser UI implementations. Replacing that model-manager layer can therefore be handled independently in a follow-up without mixing it into this UI cleanup.

User impact

  • one maintained WebUI path: the Svelte frontend embedded in audiocpp_server
  • no Python, Node.js, or Gradio is needed to run the packaged WebUI
  • model preparation may still invoke the existing Python model-manager tools
  • inference behavior and model implementations are unchanged

The cleanup removes about 16,100 obsolete lines and 48 legacy/related files.

Validation

  • npm run check — 0 errors and 0 warnings
  • npm run build — native single-file frontend rebuilt successfully
  • native language and catalog JSON files parsed successfully
  • python tools/check_loader_catalog_sync.py --self-test — passed
  • python tools/check_loader_catalog_sync.py — runtime loaders, model specs, and model manager v2 are in sync
  • Windows CUDA release audiocpp_server build — passed
  • audiocpp_server --ui --backend cpu --port 18080 smoke test — /health reported UI enabled and / returned the embedded native interface
  • git diff --check — passed

@0xShug0

0xShug0 commented Aug 13, 2026

Copy link
Copy Markdown
Owner

Thanks @mirek190! I'm building windows prebuilts for release 0.6. It seems that the prebuilts won't be affected by the cleanup so I will merge the PR after release 0.6

@mirek190

Copy link
Copy Markdown
Contributor Author

sure

I am working how to get rid of model_manager_v2.py yet

@0xShug0

0xShug0 commented Aug 13, 2026

Copy link
Copy Markdown
Owner

@mirek190
mirek190 marked this pull request as ready for review August 13, 2026 18:13
@mirek190

Copy link
Copy Markdown
Contributor Author

Fresh: https://huggingface.co/MiniMaxAI/MiniMax-Music3

sick .....
We have to replace current implementation by this one :)

@0xShug0
0xShug0 merged commit 7bad777 into 0xShug0:main Aug 13, 2026
6 checks passed
@0xShug0

0xShug0 commented Aug 13, 2026

Copy link
Copy Markdown
Owner

@mirek190 Merged! Thanks!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants