Voice Library Improvements

Is your feature request related to a problem? Please describe.
audio.cpp backend recently added support for multiple reference inputs https://github.com/0xShug0/audio.cpp/pull/282 for fish audio, and the fish speech backend it already have support for this too https://docs.fish.audio/api-reference/sdk/javascript/api-reference#ttsrequest

I think the voice library should be updated instead of having "saved voices" it should have "saved personalities". These personalities should be able to hold multiple references audio and texts and for any backend that supports multiple audio's it should include all of them when generating the TTS

Describe the solution you'd like
I'd like to be able to manage a "personality" with multiple reference audio's and texts so the output at the end can be better.

Start by reviewing the voice library and the existing Fish speech backend, then compare the multiple-reference behavior described in the audio.cpp pull request and Fish Audio TTS request documentation. Done means personalities can manage multiple reference audios and texts, and backends that support multiple audios use all of them during TTS generation.