livekit/agents: Build real-time multimodal AI applications 🤖🎙️📹

LiveKit is collaborating with OpenAI to introduce the MultimodalAgent API, enabling ultra-low latency WebRTC transport between GPT-4o and user devices. This Agents framework supports real-time AI-driven server programs that can handle text, audio, images, and video. Plugins are available for popular LLMs and services, and the framework provides features like voice agent auto-detection and telephony stack integration. The system can be run across various environments, and the core library and plugins can be easily installed via pip.