yeah exactly so you're turning speech from your user into text using a transcriber, those transcribers are set to a language, then you're sending it's output to an LLM to think about, and then you're using text to speech to give that thought a voice