Stream Voice Triggers

Stream Voice Triggers listens for spoken words or phrases and sends alerts to your OBS overlay. Each phrase can trigger a matching MP3 sound. It runs locally on your computer.

Download

Download the latest app from GitHub Releases.
Download a Vosk speech model from the Vosk models page. We recommend vosk-model-small-en-us-0.15.
Extract the Vosk model zip file to a folder you can find again, such as Documents\vosk-model-small-en-us-0.15.

Setup

Launch the Stream Voice Triggers app.
Choose the microphone or audio source you want to listen to.
Click the Vosk model chooser and select the extracted model folder, not the original zip file.
Adjust the WebSocket port or keyword triggers if needed.
Click Start.

The app starts a local WebSocket server at ws://localhost:3011 by default.

OBS Setup

Add a Browser Source in OBS.
Point the Browser Source to the local obs_overlay/index.html file.
Start Stream Voice Triggers before using the overlay.
Make sure each trigger audio value has a matching .mp3 file in the audio folder.

For example, a trigger audio value of horn needs audio/horn.mp3.