The misfolding of proteins and their aggregation in the form of fibrils or amyloid-like spherulites are involved in a spectrum of degenerative diseases. Our current understanding of protein aggregation mechanisms has primarily relied on the use of conventional spectrometric methods to determine the average growth rates and microscopic morphology of the final structures, consequently masking the morphological and growth heterogeneity of the aggregates. We developed a REal-time kinetics via binding and Photobleaching LOcalisation Microscopy (REPLOM) super-resolution method to directly observe and quantify the existence and abundance of diverse aggregate morphologies and their heterogeneous growth kinetics. Specifically, our results revealed that the growth of insulin amyloid spherulites is not exclusively isotropic but, surprisingly, may also occur anisotropically. Combining our technique with machine learning, we associated growth rates to specific morphological transitions and provided energy barriers and the energy landscape for each aggregation morphology. Our unifying framework for the detection and analysis of spherulite growth can be extended to other protein systems to reveal their aggregation processes and the broad spectrum of diverse morphologies at the single-molecule level.