The Global Burden of Disease attributes millions of premature deaths to ambient air pollution each year, making it one of the largest environmental health risks faced by society. This mortality is largely due to exposure to fine particulate matter (PM_2.5). In the United States, the Environmental Protection Agency estimated that 50.5 million people lived in counties with PM_2.5 concentrations above the level of the National Ambient Air Quality Standards in 2020. PM_2.5 levels can be derived from satellite aerosol optical depth (AOD) measurements providing comprehensive spatial and temporal coverage. However, the chemical composition of PM_2.5 affects the mechanisms by which adverse health effects occur, and thus there is a pressing need for linking satellite data with high-resolution atmospheric modeling of PM_2.5 composition. In order to better inform public health policy and decision-making, we aim to estimate near-real-time (NRT) surface PM_2.5 composition informed by satellite AOD measurements and chemical transport modeling for the first time. Here, we demonstrate this framework for hindcast estimates in year 2021. NRT AOD is collected from multi-source remote sensing data including Moderate Resolution Imaging Spectroradiometer (MODIS; Aqua and Terra), the Visible Infrared Imaging Radiometer Suite (VIIRS; Dark Target and Deep Blue), and Multi-Angle Imaging SpectroRadiometer (MISR). The data obtained from these products are combined into daily, 10-km AOD estimates and used to scale simulated total PM_2.5. GEOS-Chem (v13.1.2) nested regional simulations are run over North America with GEOS-Forward Processing (FP) assimilated meteorology at resolution 0.25° lat. x0.3125° lon. (approximately 20-30km) to simulate daily AOD and get an initial estimate of PM_2.5 composition. This estimate is interpolated into the 10-km grid and multiplied with the satellite-adjusted total PM_2.5 composition to produce concentrations of each PM_2.5 species. This satellite-constrained chemical transport model framework estimates of PM_2.5 will ultimately be evaluated against observations and compared to estimates using standard satellite products to inform future use of this framework to predict ambient air pollution health risks in true near-real-time.