ComfyUI-KokoroTTS-1038lab Introduction
ComfyUI-KokoroTTS-1038lab adds a text-to-speech node to ComfyUI. You can enter written text, choose a voice, and generate spoken audio using the Kokoro TTS framework. It can help you create narration, spoken dialogue, or voice samples as part of an AI art workflow without moving the text-generation and audio-generation steps into separate tools.
The extension supports multiple voices and languages, along with controls for speech speed and volume. English is supported by default; additional languages require configuration before they appear in the node.
How ComfyUI-KokoroTTS-1038lab Works
Think of the node as a voice performer in your ComfyUI workflow: you provide the script, select how it should sound, and the node turns the words into audio.
- Add the Kokoro TTS node to your workflow.
- Enter the text you want spoken and select an available voice.
- Optionally adjust the speech rate and volume.
- Generate the audio. To keep an audio file, connect the output to a Save Audio node.
For example, you could write a short introduction for a video, generate it in one voice, then try another voice or speech rate to see which delivery fits the scene. The node produces speech from text; it does not write or translate the script for you.
ComfyUI-KokoroTTS-1038lab Features
- Text-to-speech generation: Turn a written line into spoken audio within a ComfyUI workflow.
- Voice selection: Choose from the voices available in your configuration. Trying the same line with different voices can help you find a suitable narrator or character voice.
- Language support: The original documentation says English is available by default and describes enabling additional languages through
Languages.json. It says to remove the#from a language name in that file, then restart ComfyUI. Only enable languages for which you have the required voice and model files. - Speech-rate control: Adjust how quickly the line is delivered. A slower rate may suit deliberate narration; a faster rate may suit a more energetic passage. Listen to the result, as the best setting depends on the text and voice.
- Volume control: Adjust the level of the generated speech to better fit your intended use.
- Audio saving: Version 1.1.0 added audio-saving functionality. The documented workflow also connects the TTS output to a Save Audio node for export.
The original documentation identifies Languages.json as the place where voices are configured, but also points to config.json for a detailed voice list. If a voice you expect is missing, check the configuration files supplied with your copy of the extension rather than assuming every voice is enabled.
What's New with ComfyUI-KokoroTTS-1038lab
- Version 1.1.0 — February 24, 2025: Added audio-saving functionality. This makes it easier to retain generated speech for use outside the current workflow. See the author's version 1.1.0 update notes.
Audio-saving update screenshot
No other version changes are described in the provided documentation.
Troubleshooting ComfyUI-KokoroTTS-1038lab
A language or voice does not appear in the node
- Check whether that language is enabled in
Languages.json. The original documentation says only English is enabled by default. - Check the voice information in the configuration files included with your copy of the extension.
- Restart ComfyUI after changing language settings, then check the node again.
Speech generates, but you cannot find an exported audio file
- Check that the TTS node's audio output is connected to a Save Audio node.
- Run the workflow again.
- If you are relying on the extension's audio-saving functionality, consult the version 1.1.0 update notes for its documented behavior.
The delivery does not suit your scene
Try another available voice, then adjust speech rate and volume one setting at a time. Generate a short sample after each change so you can hear which adjustment made the difference.
Frequently asked questions
Can I use a language other than English? The author says additional languages can be enabled in Languages.json; English is the default. Restart ComfyUI after making a change.
Can I change how quickly the voice speaks? Yes. Use the node's speech-rate control and listen to a sample before generating a longer passage.
Can I save the result for another project? The documented workflow connects the node to Save Audio, and version 1.1.0 added audio-saving functionality.
Do I need a CUDA-compatible GPU? The original requirements describe one as optional for faster processing, not as a requirement.
Learn More about ComfyUI-KokoroTTS-1038lab
- Extension repository: The author's project page for the extension and its documentation.
- Version 1.1.0 update notes: Details associated with the audio-saving update.
- Kokoro TTS files: The model-file location referenced by the original documentation.
