RNA-protein interactions determine the cellular fate of RNA and are central to regulating gene expression outcomes in health and disease. To date, no method exists to identify the proteins bound to specific regions in endogenous RNAs in an unbiased fashion. Here, we develop SHIFTR (Selective RNase-H-mediated interactome framing for target RNA regions), an efficient and scalable approach to identify proteins bound to selected RNA regions in live cells. Compared to state-of-the-art RNA antisense purification techniques, SHIFTR is superior in accuracy, captures close to zero background interactions and requires orders of magnitude lower input material. We establish SHIFTR workflows for targeting RNA biotypes of different length and abundance, including short and long non-coding RNAs, as well as mRNAs and demonstrate that SHIFTR is compatible with multiplexed RNA interactome release in a single experiment. Using SHIFTR, we comprehensively identify interactions of cis-regulatory elements located at the 5ʹ and 3ʹ-terminal regions of the authentic SARS-CoV-2 RNA genome in infected cells and accurately recover known and novel interactions linked to the function of these viral RNA elements. SHIFTR enables the systematic mapping of region-resolved RNA interactomes for any RNA in any cell type, which has the potential to revolutionize our understanding of the transcriptomes of pathogens and their hosts.