jxlarrea/voice-satellite-card-integration

Voice Satellite turns any tablet, phone, or browser into a hands-free voice assistant for Home Assistant

What it solves

Voice Satellite transforms any web browser, tablet, or phone into a private, hands-free voice assistant for Home Assistant. It eliminates the need for dedicated hardware like Alexa or Google Home by allowing any device with a browser to act as a fully functional voice satellite with wake word detection and visual feedback.

How it works

The project runs as a global engine within Home Assistant that listens for wake words across all pages. It integrates directly with Home Assistant's Assist pipelines (STT, conversation agent, and TTS) and registers the browser as an assist_satellite device.

It features three on-device wake word engines running in

  • vsWakeWord: A WebGPU-accelerated engine optimized for wall-mounted tablets and far-field capture.
  • microWakeWord: A CPU-based engine that works on all devices with low latency.
  • openWakeWord: A WebGPU-accelerated engine with a broad library of pre-trained keywords.

For Android users, a companion app called Kiosk Satellite provides native CPU inference, allowing wake word detection to function even when the screen is off or other apps are in the foreground.

Who it’s for

Home Assistant users who want to add voice control to their home without relying on cloud-based assistants, specifically those using wall-mounted tablets or mobile devices as control hubs.

Highlights

  • On-device wake word detection: Supports multiple engines (WebGPU and CPU) to ensure compatibility across hardware.
  • Dual wake words: Ability to load two different wake words simultaneously, routing each to a separate Assist pipeline.
  • Rich visual feedback: Includes 9 built-in skins and a screensaver mode with support for camera feeds or media libraries.
  • Media player integration: Exposed as a TV-class device for audio, video, and camera stream playback.
  • Automation support: Provides actions for proactive announcements, starting conversations, and managing timers.
  • Experimental LLM tools: Can display visual results for weather, stocks, and web searches when paired with a tool-capable agent.

Related

  • Project
  • Project
  • Project
  • Project
  • Project