Live microphone test
Say “Hey Novo”. Audio is streamed to this server and scored every 80 ms by the same ONNX models and pipeline the iOS / Android apps run.
Setup
Models
Pick several to compare them on the same audio.
Score at or above this counts as “Hey Novo”. The apps default to 0.50.
About 2.5 s of audio around each detection is stored on this server with your name and IP address, so a person can check it and it can be used to train the model.
Input level
Microphone not started.
Not listening
Press “Start listening” and allow microphone access.