Create virtual environment and activate it

python -m venv .venv
.venv/Scripts/activate

Install Dependencies

pip install -r requirements.txt

Run KoboldAI on Google Colab

Can't pay for expensive servers so the AI is executed on Google Colab
Use "Echidna 13B (United)" (WARNING - slightly unfiltered...)
It will take around 10-25 minutes to load the model.

Paste the link from the WEB GUI

Paste the generated link into the ENDPOINT variable located in mic_vad_streaming.py
The link will look like https://crude-nevertheless-seconds-experts.trycloudflare.com/

Add .env file

Create an ElevenLabs Account and copy your API key into your .env file in the variable "ELEVEN_LABS_KEY"

Run the program

python main.py

TODO list:

1. Keep a transcript of all the things the AI said
2. Have a filter mode. This is a a barrier from KoboldAI and "elevenlabs TTS"
3. Make an interface (Vtube studio??)
4. Implement a deafen feature so the AI doesnt pick up your reactions to its input or miscellaneous noises (audio compression)