Extensions/ComfyUI-KittenTTS
ComfyUI Extension

ComfyUI-KittenTTS

A ComfyUI extension for KittenTTS

By neverbiasu·Created about a year ago·Updated about a year ago· 9
neverbiasu/ComfyUI-KittenTTS
Nodes2
On cloudLocal install
Categoryaudio/loaders, audio/tts
Stars9
Updatedabout a year ago

Nodes (2)

KittenTTS Loader
audio/loaders
KittenTTS Sampler
audio/tts
Readme

ComfyUI-KittenTTS

Lightweight text-to-speech for ComfyUI powered by Kitten TTS (ONNX). CPU-friendly, fast to start, and easy to integrate in your audio workflows.

Example Workflows

Use the included example as a reference to synthesize and save audio:

KittenTTS Workflow

Demo video:

<video src="assets/demo.mp4" controls width="720"></video>

Features

  • Lightweight ONNX model (runs well on CPU)
  • Zero-config loading (auto-downloads the default model)
  • Standard ComfyUI AUDIO output

Installation (Manual)

  1. Navigate to your ComfyUI's custom_nodes directory

    cd ComfyUI/custom_nodes

  2. Clone this repository:

    git clone https://github.com/neverbiasu/ComfyUI-KittenTTS
cd ComfyUI-KittenTTS

  3. Install Python dependencies:

    pip install -r requirements.txt

  4. Install eSpeak NG and ensure its bin folder is on PATH. Restart ComfyUI afterwards.

Troubleshooting

  • Phonemizer/eSpeak errors: install eSpeak NG, ensure it’s on PATH, then restart ComfyUI.
  • AUDIO output errors in downstream nodes: ensure upstream provides a dict with {"waveform": Tensor[B,C,T], "sample_rate": int}; this node outputs [1,1,T] at 24000 Hz.
  • Model download issues: check network access to Hugging Face and retry.

Acknowledgements

License

See the LICENSE file in this repository.