Beyond the industrial mass production of glass, glass is largely produced empirically, including in research. The production is time-consuming and not very reproducible. The chemical compositions usually rely on empirical values, empirical correlations, or trial-and-error approaches. The currently available glass databases, e.g. SciGlass and INTERGLAD, have large gaps in glass properties as they usually contain incoherent studies and/or data. In addition, glass processing parameters are usually not available. Even the original literature reference provides little information about the melting process. To solve this problem, the GlasDigital project aims to develop an ontology-based digital infrastructure for data-driven high-throughput glass development within the Plattform MaterialDigital (PMD). By connecting with the PMD Core Ontology, the Glass Ontology has been built around glass materials, manufacturing processes, literature background, and analysis methods. It also provides a common glass terminology that can be used for data extraction from literature or patents.