Post

I've been building a small voice AI app. So far, both OpenAI Realtime and Gemini Multimodal Live seem very unreliable with function calling. And Gemini insists that it did the function call😅