How can we implement multi-agent concepts with GPT-realtime?
I want to create an HR interview bot with multiple agents, each responsible for a different topic. The challenge is managing one communication stream with GPT-realtime instead of each agent having its own stream. Any ideas?
Thanks.