A non-invasive, transcutaneous, real-time viral detection device that is configured for self-administration, e.g., at a user's home. In one embodiment, and after positioning the device relative to the human body part (e.g., the user's finger), light sources in the device are activated (excited), and resulting data captured. In particular, a set of Raman spectra are collected from a configured set of emitters and detectors in the device and delivered to a nearby receiver, preferably wirelessly. The receiver filters and de-convolves the Raman spectra producing a data set representative of the constituent elements in the user's tissue of interest. The data set is applied against a statistical classifier, e.g., a neural network that has been trained to recognize and distinguish the absence or presence of viral components, e.g., C-19, or its associated blood-borne acute phase reactants. The classifier outputs an appropriate indicator, preferably in real-time, providing the user with an immediate indication of whether C-19 (or other virus of interest) is present.