Couldn't find a lightwight desktop-based OSC Headtracker so coding my own in Python. Today the maths begun working (label are wrong order video I know) but this is sending OSC to the VST on the left and controlling it through Google Mediapipe Face recognition. Obviously not full freedom, but I think the latency is good for such early days and running in the Python debugger. Next step add config file, optimise and work out what types of OSC all the other plugins might need.