WaifuAssistant

The Unified assistatnt for waifu lovers.

The goal of this project is that the girl recognize user's personality, emotion and voice. So that, in given context, the girl should response correct voice, text and face.

Demo

feature

Chatbot
TTS
ASR
VtubeStudio

WIP feature

Response with emotion
Recongnize user's emotion
Recongnize User's screen

How to run

Clone server-side repository(this repository)

git clone https://github.com/yw0nam/WaifuAssistant
cd WaifuAssistant

You should adjust port, huggingface cache path, db path.. etc in docker-compose.yaml

After that,

docker-compose build

docker-compose up

After running server side,

Clone local-side repository

git clone https://github.com/yw0nam/WaifuAssistant_local
cd WaifuAssistant_local

conda create -n WaifuAssistant_local python=3.10
conda activate WaifuAssistant_local
pip install -r requirements.txt

Install VTS desktop audio plugin by Lua Lucky
open it and connect to Vtube Studio

After that, you can run the code,

python main.py