ComfyUI-kokoro-TTS

Kokoro is an open-weight TTS model with 82 million parameters.

By Apache0ne·Created 2 years ago·Updated 2 years ago· 1

Apache0ne/ComfyUI-kokoro

Nodes2

On cloudLocal install

CategoryComfyUI jhj Kokoro Onnx

Stars1

Updated2 years ago

Nodes (2)

(Down)Load Kokoro Model

It downloads AND loads — the Kokoro model node that does both

ComfyUI jhj Kokoro Onnx

Kokoro Audio Generator

Give your ComfyUI workflow a voice — text in, WAV out, no API key

ComfyUI jhj Kokoro Onnx

Readme

ComfyUI-Kokoro TTS

This is the ComfyUI custom node wrapper for the kokoro-onnx

Dependencies

'onnxruntime-gpu' py lib
'kokoro-onnx' py lib

depen issues

kokoro-onnx will change your numpy install version

Installation

ComfyUI-Manager (rec)
Can be downloaded from ComfyUI-manager
IF using Windows Port version
ComfyUI Folder
pip install onnxruntime-gpu kokoro-onnx
Inside Custom_nodes Folder
git clone https://github.com/Apache0ne/ComfyUI-kokoro.git
IF using Matrix
Inside venv\Scripts
activate
pip install onnxruntime-gpu kokoro-onnx
Inside Custom_nodes Folder
git clone https://github.com/Apache0ne/ComfyUI-kokoro.git

Models

Models are automatically downloaded to ComfyUI\models\kokoro-onnx directory.

Updated to Latest TTS version is v1.0

Right now it uses:

Should auto download first run in workflow: IF not
models on huggingface: kokoro-models

##credits Base node built by: https://github.com/jhj0517