This teaches Claude to build production voice AI apps using the modern stack: OpenAI Realtime API for integrated voice-to-voice, Vapi for quick agent deployment, Deepgram for transcription, ElevenLabs for synthesis, and LiveKit for WebRTC infrastructure. It's opinionated about latency, pushing streaming everywhere and proper barge-in handling. The code examples are real implementations, not toy demos. Use this when you're building voice agents, real-time phone systems, or anything where users expect natural conversation flow. It won't hold your hand on audio fundamentals, but if you know the difference between PCM16 and opus, you'll move fast.
npx skills add https://github.com/davila7/claude-code-templates --skill voice-ai-development