Do any of you know of an open source local virtual assistant/companion similar to Ani from Twitter?

I am considering developing a Robo-wife via a web app which would make API calls to a home server (>=6GB GPU required)

>Robo-wife Phone Camera -> Image -> Server -> Image tagger ---(0.2 second delay)---> Image tags -> Server
>Robo-wife Phone Microphone -> Audio -> Server -> Speech to Text ---(0.2 second delay)---> Server
>Server compiles image tags and audio into a response (1 to 5 seconds) -> text to speech ---(0.4 second delay)---> Robo-wife -> Phone Speaker and Screen

Could also add in OCR to provide reading capabilities.

I think it'd be possible to get near real time responses with streaming response tokens from the server.

Can't you just with SillyTavern with TTS and a Live2D model?
Take a look at /aicg/.

>>106118326
I don't think those will work... Maybe I can use the silly tavern UI, or some other basic chat UI to avoid writing that myself

That shit is just some preset animations that play based on related keywords, its just standard chat-gpt replies that trigger those animations, how in the fuck is people losing his shit over that?

>>106118796
>>106118295 (OP)
Let's focus less on the animations - I want the same communication experience - the "animations" can be static images.

Just train an llm on lewd conversation dataset, BOOM there you have it.

>>106118817
>doesn't read the OP
nice bait, swaggot

>>106118295 (OP)
zump

>>106120164
again

>>106118295 (OP)
There is a project called open-llm-vtuber