openai-voiceassistant

A simple voice assistant with speech input and output based on openai api (chatgpt). It features:

• voice controlled: only listens on configurable keyword (assistant_name)
• voice controlled: exits listening mode on keyword
• voice controlled: enters test mode on keyword (no openai api call, just dummy text)
• voice controlled: can extend listening timeout on command

Prerequisites

Scripts expects an environment variable called OPENAI_API_KEY with an active api key to execute the api call. It can be specified in a .env file next to main script (script would fetch it automatically).

Needed python modules:

• pyaudio
• dotenv
• logging
• speech_recognition
• pyttsx3
• openai

configurations

Multiple configurations are possible at the beginning of the script:

• language (configured german)
• keywords and confirm messages
• timeouts
• used openai model

More informations can be found here: https://www.wdev.ch/how-to-create-your-own-voice-assistant-with-python-and-chatgpt-api/