Kugel Audio (documentation: https://github.com/Saganaki22/ComfyUI-KugelAudio and demonstration: https://huggingface.co/spaces/multimodalart/kugelaudio) is a text-to-speech (text-to-voice) model with 24 European voices.

Two possibilities: generation from a predefined text and language or from a cloned voice from an imported audio file.

Export wav.

Note: The demo version makes random choices of voices according to the settings and tends to accelerate at the end of the generation.

Source for the tests: Art poétique, Verlaine.