So far I haven't had any issues on my end. Using elevenlabs for the voice and gpt 4o for the model, and deepgram for the speech to text