Freeform features (FFF) manipulation is an emerging research area in geometric modelling field, which aims to develop innovative shape modification paradigms, more focused on global shape than on local curve modification. A major hindrance to FFF is given by locality properties of spline curves features representation, which forces designers to apply a direct control points manipulation, where a global modification tool or features based shaping might be more useful. Level-of-details (LOD) representation of curves, introduced by multiresolution wavelet analysis (MRA), allows to overcome spline locality weakness by describing features as detail coefficients, which can be reused over other different curves. In this paper we present a global modelling tool, based on MRA properties, useful to capture and extract curve features. Features, stored as a simple set of coefficients, may be easily applied over other curves, operating a smart modelling action and returning a goodlooking merge bet...