Is possible to intercept speech to text recognition as text?

Hi, is possible to intercept speech to text recognition? My goal would be talk with muse luxe and write my word in a template sensor for do some test with chatgpt.