A local macOS voice-control integration for OpenClaw that enables wakeword-triggered AI interactions with ASR, TTS, and overlay UI support.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install openclaw-voice-control
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install openclaw-voice-control using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is OpenClaw Voice Control?
OpenClaw Voice Control provides a seamless, local-first bridge between voice input and your AI agent. Designed specifically for macOS, this integration brings high-performance Automatic Speech Recognition (ASR) via FunASR and SenseVoice, coupled with customizable Text-to-Speech (TTS). It allows developers and power users to interact with Openclaw Skills without touching a keyboard, using efficient background services managed by launchd.
This skill is built with privacy and performance in mind, keeping most processing local to the machine. By integrating an optional overlay UI and robust wakeword activation (supporting both openWakeWord and Porcupine), it transforms OpenClaw from a text-based interface into a fully interactive voice assistant that integrates deeply with the macOS ecosystem.
OpenClaw Voice Control Use Cases
- Hands-free coding and system operation via voice commands to local AI agents.
- Triggering complex workflows using the "hey jarvis" wakeword for seamless Openclaw Skills execution.
- Real-time voice interaction with visual feedback through the optional overlay UI.
- Setting up background-resident AI assistants that start automatically upon macOS login.
How OpenClaw Voice Control Works
- The system listens for a specific wakeword using the local microphone and the openWakeWord or Porcupine provider.
- Once triggered, the SenseVoice engine captures and transcribes the user speech into text locally on the Mac.
- The recognized text is forwarded to a local OpenClaw agent for processing and response generation.
- The agent response is converted back to speech using macOS TTS and played through system speakers.
- An optional overlay UI provides visual status and text feedback during the conversation loop.
- For persistent use, launchd manages the lifecycle of the service as a background resident process.
OpenClaw Voice Control Setup
# From your conversation agent's installed skill directory
git init
git remote add origin https://github.com/CarrotYuan/openclaw-voice-control.git
git fetch --depth 1 origin main
git checkout -B main FETCH_HEAD
# Setup environment and dependencies
python3 -m venv .venv
source .venv/bin/activate
pip install -e .
# Download ASR and VAD models
./.venv/bin/modelscope download --model iic/SenseVoiceSmall --local_dir models/SenseVoiceSmall
./.venv/bin/python - <<'PY'
from funasr import AutoModel
AutoModel(model='fsmn-vad', disable_update=True)
PY
# Configure environment
cp .env.example .env
# Update .env with your OPENCLAW_BASE_URL and OPENCLAW_TOKEN
OpenClaw Voice Control Data Schema & Taxonomy
| Component | Description |
|---|---|
.env |
Stores environment variables like OPENCLAW_TOKEN and model paths. |
models/ |
Local directory containing SenseVoice and VAD model weights. |
config/default.yaml |
Main configuration file for tuning wakeword thresholds and timing. |
launchagents/ |
Metadata and plist files for macOS launchd background service management. |
scripts/ |
Contains maintenance tools like doctor.sh and deployment wrappers. |
OpenClaw Voice Control Advanced Features
- Background resident behavior using launchd for persistent, always-on voice access to Openclaw Skills.
- Support for multiple wakeword engines including openWakeWord (built-in models) and Picovoice Porcupine (custom .ppn files).
- Integrated diagnostic tools through the doctor.sh script to validate environment and dependency health.
- Automatic user-login startup support via the provided macOS deployment scripts.
- Visual overlay application that synchronizes with the voice service to provide UI feedback.
SKILL.md
Loading
Related Openclaw Skills
Domain Recommender
A high-performance utility for generating brandable, SEO-optimized domain names with integrated live availability verification for AI products.
Birth Chart Analysis
An advanced astrological computation engine that generates precise natal charts, planetary positions, and house placements from birth data.
Showcase Page Generator
A specialized technical workflow for generating, optimizing, and auditing showcase pages that highlight user-generated content and social proof.
Halo CMS Skill
A professional blog management skill for the Halo CMS that enables automated post creation, comment replies, and taxonomy management via REST API.
Services Page Generator
A professional framework for AI agents to architect, write, and optimize service-based business pages for agencies, consultants, and freelancers.
AI Persona OS
AI Persona OS is a comprehensive operating system for Openclaw Skills agents that provides robust personality management, memory persistence, and context protection protocols.
METADATA
Featured*
