Feature request: Allow for combination of voices when using kokoro fast api

[artaeun]

Would be great if it allowed to combine voices with kokoro fast api. It's should be a matter of just passing to "voice1(weight1)+voice2(weight2)" to the api instead of "voice1(weight1)", in theory, but I have no idea how hard it is to implement

[jukofyork]

Yeah, this would be a great addition. I've tried quite a few combinations and bf_emma+af_heart seems the most natural sounding to me (as someone from the UK).

Using Kokoro-FastAPI specifically, it's simply a case of passing the names using the plus syntax:

https://github.com/remsky/Kokoro-FastAPI?tab=readme-ov-file#features

I'm not sure about other APIs though, and the current drop-down selection method might make it harder to add.