Detect wake words and stream LLM responses with VOICEVOX TTS.
npx skills add https://github.com/kentoku24/voiceclaw --skill voiceclaw-jp
Please help me install this Agent Skill. Skill: voiceclaw-jp Source: https://github.com/kentoku24/voiceclaw/tree/main Command: npx skills add https://github.com/kentoku24/voiceclaw --skill voiceclaw-jp
This skill enables hands-free voice interaction with an OpenClaw agent by wake-word detection, streaming LLM responses, and text-to-speech, eliminating the need to type commands.
Clone the repo, install dependencies, and start the server to run voiceclaw locally.
High-intent search queries and answers about installing and using this skill.
Wake-word detection enables always-listening voice activation so you can trigger an LLM agent without pressing buttons. It detects a configurable phrase locally, then initiates speech-to-text capture and streams the command to the LLM for processing.
Yes, VOICEVOX supports Japanese text-to-speech for local voice assistants. You need a local VOICEVOX endpoint to synthesize high-quality Japanese speech from the streaming LLM responses generated by the OpenClaw agent.
Running a wake-word voice assistant locally requires Node.js 18+, a local OpenClaw gateway, and a VOICEVOX endpoint for TTS. A browser interface supporting the Web Speech API is also needed to handle speech-to-text input.
To reduce latency when streaming LLM responses to text-to-speech, process text at the sentence level. Streaming sentence-level LLM responses allows the TTS engine to begin audio synthesis before the full generation completes, enabling low-latency dialogue.
youngkentInteract with a self-hosted OpenClaw gateway via a local web interface.
markus41Integrate local LLMs with Home Assistant using Ollama and LocalAI.
tangzheng202202Route wake-word voice commands to multiple AI agents via HTTP adapters.
shreed27Detect wake words, transcribe speech, and provide spoken feedback for trading applications.
