Abstract The engineering of human ARRDC1-mediated microvesicles (ARMMs) as new non-viral vehicles for delivery of gene therapies overcomes challenges associated with current modalities. Hurdles such as generating sufficient material to meet demand and development of appropriate characterization assays, however, persist. Our study evaluated two scalable strategies to generate GFP-loaded ARMMs, transient transfection or stable cell line-based production. The upstream ARMMs production processes utilized a suspension HEK293-derived line, termed 5B8, from Lonza. Production was evaluated in shake flasks and bioreactors. Downstream ARMMs purification processes employed Tangential Flow Filtration (TFF) and Anion Exchange Chromatography (AEX). Analytical methods included single particle analysis, ELISA, and immunoblotting. Additionally, an in vivo study was conducted in mice to investigate the half-life and biodistribution of ARMMs administered intravenously. 5B8 cells yielded robust production of ARMMs after transient transfection with the ARMMs loading construct or using a stable cell line containing a transgene that encodes the ARMMs loading cassette, in shake flasks or a stirred tank bioreactor, respectively. Approximately 50% of all vesicles produced were payload-containing ARMMs. ARMMs were purified by ultracentrifugation (small scale) or a combination of TFF and AEX (large scale). Both purification methods produced comparable ARMMs. In vivo , ARMMs showed rapid biodistribution predominantly to the spleen and liver and, to a lesser extent, kidneys, and lungs. The successful scale-up of ARMMs production illustrates the potential of engineered extracellular vesicles (EVs). Furthermore, this study highlights the potential utility of ARMMs for in vivo delivery of therapeutic molecules.