Quantitative knowledge of liquid viscosity is of fundamental importance in many areas of materials synthesis and processing. However, the determination of viscosity often relies on specialized experimental equipment, offline experimentation, or invasive procedures, in particular when required in extreme conditions such as at high temperature, high pressure, and in confined or corrosive environments. Here, this study proposes and validates a fast and simple method that mimics the intuitive perception of liquid flow within a quantitative framework. For this, this study employs digital video observation to derive quantitative values of the shear viscosity of liquids, with high precision and rapid acquisition rates. The technique involves capturing liquid dynamics after minor mechanical stimulation. Processed imaging data are indexed by similarity and referenced to a digital database generated with a finite element model, from which values of viscosity are obtained in line. The approach is tested on water at room temperature and on a high-temperature glass melt. Covering a viscosity range of four orders of magnitude, both yield convincing agreement with tabulated reference data at low computational cost.