Putting this as a stable diffusion file, i think civitai needs to expand to audio.
Record your voice with a microphone saying a catch phrase (15 seconds recording or less is all you need)
type out what you want the ai to say in your voice and click que
play the result