We report on a computer algorithm for the in situ analysis of reflectance anisotropy (RA) spectra in a time frame compatible with the epitaxial growth of cubic semiconductors. This algorithm allows for the in situ acquisition of RA spectra and their decomposition into two components whose amplitude depends on the As coverage of the semiconductor surface. One of such components is associated with the surface orthorhombic strain due to the surface reconstruction and has an amplitude that strongly depends with surface reconstruction and thus As coverage. This fact opens the possibility of using reflectance anisotropy spectroscopy (RAS) as an optical probe to characterize the As surface coverage in real time. To demonstrate the performance of the algorithm we report on RA measurements carried out during the homoepitaxial growth of GaAs (001). We show that the algorithm is capable of analyzing a set of 500 RA spectra in a time span of about 10 s. This allows for a range of applications for the developed algorithm, including the surface characterization and fine tuning of the substrate stoichiometry just before epitaxial growth, during the growth of the buffer layer.