The mechanical fracture compliance is of interest in a number of geoscientific applications. Seismic borehole methods, especially full-waveform sonic (FWS) data, have indicated their potential to infer the compliance of macroscopic fractures under in situ conditions. These approaches rely on the assumption of a homogeneous background embedding the fractures and, as of yet, compliance estimates for individual fractures are limited to static FWS measurements. In this work, we assess the potential of inferring the compliance of individual fractures from standard, production-type FWS data in the presence of background heterogeneity. We first perform a comparative test on synthetic data to evaluate three approaches known as the transmission, phase, and group time delay methods. The results indicate that the former two produce adequate compliance estimates for scenarios with a strongly heterogeneous background or a damage zone around the fracture. These two methods are then applied to two FWS data sets acquired before and after a hydraulic stimulation campaign in a crystalline rock, which allows to test them on natural and man-made fractures. The transmission method turned out to be unsuitable for the considered data due to its reliance on amplitudes. Conversely, the travel time behavior remained stable and the phase time delay method produced robust and consistent estimates. The results for a newly created hydro-fracture imply the capability of resolving remarkably small compliance values of the order of 10−14 m/Pa. This estimate is one order-of-magnitude smaller than that for the natural fracture, which may help to distinguish between these two fracture types.