Is SOFA time-aware?

[theartisan1]

I have a project I'm trying to replace other aligners (MFA, etc) with SOFA - but the best I can get out of it seems to be duration only. I can't get it to act like it's time-aware (all words will be aligned mostly in the silence before the vocals actually begin. Is SOFA able to determine actual start and end times for phonemes, or essentially just designed to determine durations?
*Edit: The source audio is vocal-only wav