Does your model support real-time semantic end-of-speech detection for audio?

I’m currently using VAD, but sometimes if the user speaks for too long (with no silence in between), it fails to detect the end.
Is it possible for me to use your model to perform real-time semantic stop prediction?