TTS and STT. Uses an OpenAI compatible server (vLLM, OpenAI, https://github.com/njelicic/openai-server, etc.) to generate the LLM response.
git clone https://github.com/erasmus-university-library/digiderius
cd digiderius
pip3 install -r requirements.txt
cd digiderius
mkdir tts_models
wget -i ../downloads.txt
python3 cli.py \
--host=127.0.0.1 \
--port=8080 \
--OPENAI_API_URL="http://127.0.0.1:8000/v1" \
--OPENAI_API_KEY="sk-my-secret-key"
Next, navigate to http://127.0.0.1:8080/ and start talking.