Comment by TuringNYC

We use this exact stack at work (OpenAI, ElevenLabs, Deepgram) for some exploratory use cases. The key issue we have now is latency with the LLM. Deepgram and Elevanlabs work brilliantly!