SHAMAN is a computational method for identifying potential small-molecule binding sites in RNA structural ensembles. It combines atomistic molecular dynamics simulations with enhanced-sampling techniques to explore RNA conformational landscapes and efficiently identify RNA pockets. SHAMAN was benchmarked on a set of biologically relevant RNA targets, including large, structured riboswitches and small, flexible viral RNAs. It successfully identified all experimentally resolved pockets and ranked them among the most favorable probe hotspots. SHAMAN addresses the challenges of RNA dynamics by considering both local and global flexibility, making it a valuable tool for rational drug design targeting RNA with small molecules. RNA molecules, once thought to be only carriers of genetic information, now perform diverse biological functions, including regulating protein synthesis and defending against foreign nucleic acids. Targeting RNA with small molecules is emerging as a promising strategy for treating diseases such as cancer, viral infections, and neurodegenerative conditions. However, the dynamic nature of RNA and limited understanding of its structural and dynamic properties hinder rational drug design. SHAMAN overcomes these challenges by using small probes and metadynamics-enhanced sampling to explore RNA conformational changes and identify binding sites. SHAMAN was tested on 7 biologically relevant systems, including riboswitches and viral RNAs. It successfully identified all experimentally resolved binding sites and ranked them among the most probable SHAMAPs. The accuracy of SHAMAN was evaluated by comparing the distance between SHAMAPs and experimental binding sites. The results showed that SHAMAN's SHAMAPs were closely aligned with experimental binding sites, with an average distance of 3.8 Å and 4.4 Å for holo-like and apo cases, respectively. The probes used in SHAMAN simulations were analyzed to understand their relation to physico-chemical features of RNA pockets and small molecules. Aromatic probes showed a preference for exploring deep cavities, while non-aromatic probes explored shallow pockets. SHAMAN was compared with other tools for small-molecule binding site prediction, and it outperformed them in terms of accuracy and precision. The method was tested on two case studies: the FMN riboswitch and the HIV-1 TAR element. SHAMAN successfully identified both binding modes of the FMN riboswitch and the cryptic binding pocket in the HIV-1 TAR element. SHAMAN is a promising approach for identifying small-molecule binding sites in RNA structural ensembles. It provides a foundation for future drug design efforts targeting RNA with small molecules, effectively addressing the long-standing challenges in the field. The accuracy, reliability, and versatility of SHAMAN in identifying small-molecule binding sites across diverse RNA systems highlight its potential value in the field. By integrating SHAMAN in virtual screening pipelines, we aim to create an advanced platform for the rationalization of RNA-targeting molecules.SHAMAN is a computational method for identifying potential small-molecule binding sites in RNA structural ensembles. It combines atomistic molecular dynamics simulations with enhanced-sampling techniques to explore RNA conformational landscapes and efficiently identify RNA pockets. SHAMAN was benchmarked on a set of biologically relevant RNA targets, including large, structured riboswitches and small, flexible viral RNAs. It successfully identified all experimentally resolved pockets and ranked them among the most favorable probe hotspots. SHAMAN addresses the challenges of RNA dynamics by considering both local and global flexibility, making it a valuable tool for rational drug design targeting RNA with small molecules. RNA molecules, once thought to be only carriers of genetic information, now perform diverse biological functions, including regulating protein synthesis and defending against foreign nucleic acids. Targeting RNA with small molecules is emerging as a promising strategy for treating diseases such as cancer, viral infections, and neurodegenerative conditions. However, the dynamic nature of RNA and limited understanding of its structural and dynamic properties hinder rational drug design. SHAMAN overcomes these challenges by using small probes and metadynamics-enhanced sampling to explore RNA conformational changes and identify binding sites. SHAMAN was tested on 7 biologically relevant systems, including riboswitches and viral RNAs. It successfully identified all experimentally resolved binding sites and ranked them among the most probable SHAMAPs. The accuracy of SHAMAN was evaluated by comparing the distance between SHAMAPs and experimental binding sites. The results showed that SHAMAN's SHAMAPs were closely aligned with experimental binding sites, with an average distance of 3.8 Å and 4.4 Å for holo-like and apo cases, respectively. The probes used in SHAMAN simulations were analyzed to understand their relation to physico-chemical features of RNA pockets and small molecules. Aromatic probes showed a preference for exploring deep cavities, while non-aromatic probes explored shallow pockets. SHAMAN was compared with other tools for small-molecule binding site prediction, and it outperformed them in terms of accuracy and precision. The method was tested on two case studies: the FMN riboswitch and the HIV-1 TAR element. SHAMAN successfully identified both binding modes of the FMN riboswitch and the cryptic binding pocket in the HIV-1 TAR element. SHAMAN is a promising approach for identifying small-molecule binding sites in RNA structural ensembles. It provides a foundation for future drug design efforts targeting RNA with small molecules, effectively addressing the long-standing challenges in the field. The accuracy, reliability, and versatility of SHAMAN in identifying small-molecule binding sites across diverse RNA systems highlight its potential value in the field. By integrating SHAMAN in virtual screening pipelines, we aim to create an advanced platform for the rationalization of RNA-targeting molecules.