Exosomes, which are membranous nanovesicles, are actively released by cells and have been attributed to roles in cell-cell communication, cancer metastasis, and early disease diagnostics. The small size (30-100 nm) along with low refractive index contrast of exosomes makes direct characterization and phenotypical classification very difficult. In this work we present a method based on Single Particle Interferometric Reflectance Imaging Sensor (SP-IRIS) that allows multiplexed phenotyping and digital counting of various populations of individual exosomes (>50 nm) captured on a microarray-based solid phase chip. We demonstrate these characterization concepts using purified exosomes from a HEK 293 cell culture. As a demonstration of clinical utility, we characterize exosomes directly from human cerebrospinal fluid (hCSF). Our interferometric imaging method could capture, from a very small hCSF volume (20 uL), nanoparticles that have a size compatible with exosomes, using antibodies directed against tetraspanins. With this unprecedented capability, we foresee revolutionary implications in the clinical field with improvements in diagnosis and stratification of patients affected by different disorders.