Thanks — I’m responding to your request for recommendations on an agentic AI framework to use with LiveKit support. Short answer • Use LiveKit Agents. It’s an open-source framework built specifically to create agentic AI applications that participate in LiveKit rooms (voice, video, and text) and it natively integrates with many AI providers and tooling. Why LiveKit Agents is a strong fit • Real-time AI pipelines: Supports streaming STT → LLM → TTS with reliable turn detection and interruption handling, which is essential for live support conversations. • Multi-agent workflows: You can decompose complex support flows into discrete agent steps and coordinate closed-loop workflows. • Plugin-based integrations: Over 40 plugins already available for LLMs (OpenAI, Anthropic, Google, Mistral, Azure, etc.), STT/TTS providers, VAD, avatars, and tools — so you can mix-and-match providers that suit your latency, cost, or fidelity requirements. • SDKs and languages: Built examples and SDKs in Python and Node.js, making it easy to iterate and integrate with most stacks. • Production readiness: Includes job dispatch, load balancing, automatic failover, and scale options; the same agent code can run locally and be deployed at scale (including LiveKit Cloud Agents). Common use cases you can build quickly • Live audio/text customer support agents that join rooms and assist human agents or handle basic inquiries. • AI-assisted call centers / telephony (SIP) integrations for inbound/outbound voice. • Multimodal agents combining avatar video, TTS, and LLM-driven dialog for guided support flows. If you want next steps, tell me: • Which AI providers you prefer (OpenAI, Anthropic, Azure, etc.), or whether cost/latency is a priority. • Whether you need telephony (SIP) support or just in-app voice/video. • Your preferred language (Python or Node.js) and whether you want sample code or an architecture sketch. I can then suggest a concrete integration plan (components, plugins to use, and a small starter checklist) or provide sample code for the stack you pick. Sources: Agents (Python) | livekit/livekit_composite | DeepWiki | Introduction | LiveKit Documentation | Building the all-in-one platform for voice AI agents | An open source stack for real-time multimodal AI | OpenAI and LiveKit | LiveKit Documentation