endman100-ComfyUI-HiggsAudioV3TTS/examples

Higgs Audio V3 TTS Example Workflows

These examples assume:

• ComfyUI is running at http://127.0.0.1:8188.
• Higgs Audio V3 Model Loader can start a local Higgs runtime. If SGLang-Omni is installed outside ComfyUI's Python environment, set HIGGS_AUDIO_V3_PYTHON_EXECUTABLE before starting ComfyUI.
• Every JSON file in this folder is a drag-and-drop ComfyUI UI workflow. API /prompt JSON is intentionally not stored here because it does not load as nodes when dragged onto the canvas.

Files

• higgs_audio_v3_tts_local_model_workflow.json

  • Minimal no-server workflow: Higgs Audio V3 Model Loader -> Higgs Audio V3 Local TTS -> SaveAudio.

• higgs_audio_v3_tts_local_features_workflow.json

  • Drag this into the ComfyUI UI to test local model loading, basic TTS, generated AUDIO reference input, multilingual text, and inline Higgs control tokens.

Expected Output

The workflows write audio files under:

<ComfyUI output>/higgs_audio_v3_tts/