Capitalizing on advancements in Large Language Models (LLMs), MetaSupra aspires to expedite the process of metadata enrichment in FAIR-compliant repositories.
Specifically, MetaSupra will enhance SupraBank, a platform that provides machinereadable physicochemical parameters and metadata of intermolecular interactions in solution. By utilizing LLMs, we aim to develop data crawlers capable of extracting context-specific information from chemical literature, simplifying data acquisition, and curating FAIR repositories. This crawler software will be made accessible, inviting potential adoption by other Helmholtz centers. In addition, MetaSupra will illustrate how repositories' utility for correlation studies, machine learning, and educational purposes can be substantially amplified through the integration of quantum-chemically computed molecular parameters, positioning it as a model for other chemical repositories and moving forward with its integration into IUPAC activitie