Extracellular vesicles (EVs) have prevalent roles in cancer biology and regenerative medicine. Conventional techniques for characterising EVs including electron microscopy (EM), nanoparticle tracking analysis (NTA) and tuneable resistive pulse sensing (TRPS), have been reported to produce high variability in particle count (EM) and poor sensitivity in detecting EVs below 50 nm in size (NTA and TRPS), making accurate and unbiased EV analysis technically challenging. This study introduces direct stochastic optical reconstruction microscopy (d-STORM) as an efficient and reliable characterisation approach for stem cell-derived EVs. Using a photo-switchable lipid dye, d-STORM imaging enabled rapid detection of EVs down to 20–30 nm in size with higher sensitivity and lower variability compared to EM, NTA and TRPS techniques. Imaging of EV uptake by live stem cells in culture further confirmed the potential of this approach for downstream cell biology applications and for the analysis of vesicle-based cell-cell communication.