Hey everyone! Sharing the first version of my EchoMuse-to-Hubitat voice bridge.
It connects EchoMuse to Hubitat for voice-controlled lights and switches, using Whisper for speech recognition, an OpenAI-compatible LLM for interpreting requests, and Piper for spoken replies. No Home Assistant required. Supports on/off commands, brightness percentages, and a default speaker room.
Full disclosure: this was completely vibe coded. Very few brain cells were actually used in this process, and the remaining ones are requesting time off.
This is an early experimental release, not a polished integration. Start in dry-run mode, only allow devices you’re comfortable experimenting with, and expect bugs and rough edges.
Code and setup instructions:
Feedback, fixes, and contributions from people who actually know what they’re doing are very welcome!
Update, October 27: Added experimental Hubitat-to-EchoMuse announcements. See the update below for installation details and testing status.
Added support for Hubitat to send spoken announcements to EchoMuse without a wake word.
The updated bridge app automatically creates an EchoMuse Speaker device using a companion driver. You can use it in Hubitat notifications and automations for things like “The front door has opened.”
Fixed announcements go straight through Piper, so no LLM is involved.
The direct announcement test is working on my setup. The new Hubitat app/driver workflow still needs end-to-end testing, so this is still early alpha.
For anyone updating: install the new speaker driver first, update the existing bridge app and Python code, then configure the announcement settings in the app.
The README has been updated with setup instructions. Feedback and testing welcome!