That project is using the acceleration due to gravity to provide a local "upwards" vector.
This isn't going to work in a rocket as the dominant acceleration will be due to the rocket's propulsion system. In fact most accelerometers go to +/-8g and most model rockets will exceed that substantially at times I reckon!
So all you will be doing is measuring the angle between the rocket's thrust axis and the accelerometer's axis.
To do the kind of stabilization you are wanting you need rate gyros - these shouldn't be (much) affected by linear acceleration but will detect rotation rates. There are lots of resources out there about flight stabilization using 3, 6 and 9 DoF inertial sensing. Look for these.
Silicon MEMS rate gyros have significant drift, but on a timescale of 10's of seconds, probably not a worry for a rocket!!