Single-cell methodologies have improved our understanding of cellular diversity within complex tissues such as the mammalian neocortex. However, despite the unprecedented granularity, sampling of minority yet diverse cell populations remains challenging. Here, we show that many subtypes of cortical inhibitory γ-aminobutyric acid-containing (GABAergic) neurons are poorly represented in current atlases, with most datasets limited to adult stages and not recapitulating cell-state transitions of these rare cell types during development. We address this gap by optimizing experimental techniques and computational pipelines to enrich and enhance the resolution and representation of these cell types. Leveraging our previously published scRNAseq datasets, we developed a computational pipeline that integrates additional cel​ls from other developmental stages, resulting in a comprehensive atlas of SST+ neurons – one of the most diverse classes of inhibitory neurons. By optimizing clustering parameters and filtering unstable clusters and cells, we achieved a high-resolution atlas comprising more than 40,000 mouse cortical SST+ neurons, over 4 developmental stages, with highly stable cell clusters (Jaccard index >0.8). To validate our pipelines, we generated a parallel atlas composing of PV+ interneurons. Finally, we developed a web-based tool for continuous improvement and integration of new datasets, ensuring accessibility and usability for researchers with varying computational resources. In summary, our study developed two cortical inhibitory developmental atlases that are continuously updated, easily accessible through a web-based interface, and map cell identity queries from data generated on different sequencing platforms, offering insights into the diversity of inhibitory cell identities during development.