Does anyone here have a model of how to use groq to transcribe audio. I've tried everything, I can do it through n8n, but I can't do it directly through typebot. Typebot generates an audio file in a .webm url that openai can transcribe without any problems. However, I can't do this through groq. Currently I use a flow in typebot, convert it in n8n and return it to typebot, but I know there must be a way to convert it directly through typebot.