You're asking whether LiveKit has done work on device-directed speech detection (Apple-style Voice Trigger). LiveKit does not offer a built-in device-directed speech / wake-word (Voice Trigger) system. LiveKit provides voice activity detection (VAD), semantic end-of-turn detection, and transcription pipelines, and you can implement keyword detection in STT streams or custom pipeline nodes, but there is no native device-directed speech detection feature. If you'd like, I can: • Outline how to build a wake-word or device-directed flow using LiveKit's VAD/turn detection plus an STT or keyword detector, • Suggest which pipeline nodes/hooks to use for keyword detection, • Or ask a few questions about your target devices, latency, and accuracy needs to recommend an approach. Which would you prefer? Sources: Improving voice AI's turn detection with transformers | LiveKit turn detector plugin | LiveKit docs | An open source stack for real-time multimodal AI | Pipeline nodes and hooks | LiveKit docs