image

• Chat(OpenAI powered by azure cognitive services)
• Speech(powered by azure cognitive services)
• OCR by cnOCR (powerd by paddlepaddleOCR)

for localhost solution, speech to text and text to speech can be instead by:

• Bark(Text to Speech)
• Whisper(Speech to Text)

brew install libespeak1 espeak libspnav swig portaudio sox
virtualenv --python=python3.10 .venv
source .venv/bin/activate
pip install -r requirements.txt
python -m spacy download zh_core_web_sm en_core_web_sm

Edit your template file and replace api keys & models

cp .env.template .env
python app.py

also you need to setting a magic word for converstion, for example. i use 下次再聊吧 to end the Converstion.