@chassain-0
Some of the voice models of snowboy sometimes looks weird. Jarvis is one example.
They put the 2 voice patterns in one model(“jarvis.umdl”), so you should use like this;
https://github.com/eouia/MMM-Hotword/blob/master/documents/models.md
0_1577961062991_3e29c14b-02d1-4249-a926-b741b5832f4a-image.png
Or you can assign a different identifier to each pattern like this;
{
hotwords: ["jarvis_1", "jarvis_2"],
file: "jarvis.umdl",
sensitivity: "0.8,0.8",
},