I have a pre-made RVC voice model that I want to use as a custom voice for an AI assistant/TTS.

[rugged spade]

How the heck do I accomplish that in Python? (I'm a complete newbie so I'm still learning.)

From what I understand it involves taking a TTS voice, converting it to the desired voice using RVC/the path and index files, and outputting the new voice. Obviously it's probably more complex than that. I'm working towards building my own simple AI assistant with a custom voice. I've used Applio and whatnot before. But I'm just unsure how to accomplish what I'd like to accomplish.

Basically, I have the voice of a game character that I'd like to use as the TTS voice.

[fast sleet]

what's your pc gpu and operative system? I'm guessing you already have chosen an LLM to use?

[rugged spade]

3080 ti and Windows 11, in terms of LLM, OpenAI is probably the "best" option isn't it? Which obviously isn't free. So I might end up going the free route until I'm competent enough with programming. I'll need to look more into the free options.

[fast sleet]

you got an rtx 3080 ti 12gb vram

you can even use models locally, even tho smaller to models like gemini 2.5 pro or chatgpt, but open source models that are completely free and limitless (unless you count your hardware) like gemma 3

or if you want to use an api that has free options for the best models, there's https://groq.com/ or https://aistudio.google.com

[rugged spade]

How do I go about the whole RVC model ordeal? Is using one of them for TTS difficult to set up?

[fast sleet]

RVC is Speech To Speech Natively

not TTS

you first of all need to get an LLM via maybe llama.cpp locally or an API

then use any type of TTS you want, the easiest is edge tts which is free and cloud, multilingual and good quality but without emotions https://pypi.org/project/edge-tts/

[rugged spade]

Noted, long way to go in terms of learning either way. Thanks