I am using Deepgram to transcribe the live call, consume the result, and feed it to LLM. I have set the interim results to false, but Deepgram sends multiple results with is_final=True, one every 5 seconds.
Is there any way to control/modify this behaviour? Instead of 5 seconds, can I configure it for 10 seconds?
Is there any way to identify that this is not the final transcript? Can I rely on speech_final parameter and aggregate the results till I don't receive speech_final param?