Provides continuous conversation with microphone and speakers like using Google Assistant After starting the script you'll be asked to resolve the captcha in the puppeteer browser instance the script will then listen to your default microphone until silence of about 5 seconds has been reached, the language model will then try to guess the things you said and pass them on to chatGPT and its output will be used to fetch googleTTS files to play on the default system speaker
- Create
.env
fileOPENAI_EMAIL="your_email" OPENAI_PASSWORD="your_password" CHROME_PATH="your_chrome_path"
npm install
- install
sox
(if on windows) and libmad-0.dll for sox - install preferred
vosk
models for speech recognition, put them intomodel/
folder - resolve other missing dependencies
npm run dev