ComfyUI-Sopro

Sopro TTS custom nodes for ComfyUI - Lightweight CPU-based text-to-speech with zero-shot voice cloning.

Features

• ⚡ Runs efficiently on CPU (0.25 RTF - 30s audio in 7.5s)
• 🎤 Zero-shot voice cloning with 3-12s reference audio
• 🔧 Compatible with all ComfyUI audio workflows
• 💾 169M parameters - lightweight and fast

Installation

1. Navigate to your ComfyUI custom_nodes directory:

cd ComfyUI/custom_nodes/

2. Clone this repository:

git clone https://github.com/ai-joe-git/ComfyUI-Sopro.git
cd ComfyUI-Sopro

3. Install dependencies:

pip install -r requirements.txt

4. Restart ComfyUI

Nodes

Sopro TTS Generator

Main text-to-speech generation node with optional voice cloning.

Inputs:

• text (required): Text to synthesize
• reference_audio (optional): Reference audio for voice cloning
• speed (optional): Speech speed (0.5-2.0, default 1.0)
• temperature (optional): Generation temperature (0.1-1.5, default 0.7)
• seed (optional): Random seed for reproducibility

Outputs:

• audio: Generated audio in ComfyUI format

Sopro Load Reference Audio

Load audio files for voice cloning.

Inputs:

• audio_file: Audio file from input directory

Outputs:

• reference_audio: Audio in ComfyUI format

Sopro Save Audio

Save generated audio to output directory.

Inputs:

• audio: Audio to save
• filename_prefix: Output filename prefix
• format: Output format (wav/mp3/flac)

Example Workflow

1. Add "Sopro TTS Generator" node
2. Enter your text
3. (Optional) Add "Sopro Load Reference Audio" and connect for voice cloning
4. Connect output to "Sopro Save Audio" or other audio processing nodes
5. Generate!

Tips

• Use phonemes instead of abbreviations (e.g., "1 plus 2" not "1 + 2")
• Reference audio should be 3-12 seconds for best voice cloning
• Lower temperature for more consistent output
• Works great with other ComfyUI audio nodes!

Credits

• Sopro TTS by samuel-vitorino