Hey, great job here! Two questions:
1) Is it at all feasible to do the stabilisation on the fly so that the screen shows already-stabilised video (if not then why)?
2) Do you envisage building an API so that a third-party app can use your technology?