pyttsx3 generates robotic voice which is not even understandable no matter how I tried to change language, speed and tone.
Thus, gtts is used and saved as speech.mp3, which is again played by playsound module.
The script is run in the background, listening to any dialog.
Once the hot-word or trigger-word, JARVIS is heard, it will start responding.
set_alarm function is threaded so that other functions and commands can still be used while alarm is scheduling in background.
GPT-3.5 is utilized as assistant bot.
- Jarvis play (music)
- Jarvis what's the time
- Jarvis set alarm --:-- AM/PM
- stop/turn off alarm
- Jarvis what's the temperature
- Jarvis how's the weather
- Jarvis save memo
- Jarvis show me the list of memo
$ git clone https://github.com/leo007-htun/GPT3_Automated_Assistant_with_wake_word.git
$ pip install -r requirement.txt
replace YOUR_API_KEY with user's OpeanAI API key
$ source sr.sh