Recently I participated Monday ai Hackathon where I combined the Google text to speech model(you can also use whisper by open ai )and Codex model to create a dynamic solution that enables voice commands to manipulate database tables and automate SQL query generation

To give voice commands first you had to record voice which we can easily by installing Wavio and sounddevice libraries of Python for saving audio files with timestamps you can import the time library

To convert speech into text from google cloud api first we ,need to write code to extract the transcript from API response. Remember to match the audio attributes for record and transcript functions.

and then we have to pass that transcript into the codex model to generate sql queries .you can also add your table name and attributes of table with which you want this model to connect in the prompt to be more specific

you can find the project code here https://github.com/MN-Noor/voice-assistant-