WSInsight as a cloud-native pipeline for single-cell pathology inference on whole-slide images
WSInsight is an open-source platform for cohort-scale H&E whole-slide image analysis. It performs patch-level inference together with single-cell segmentation and phenotype classification, with morphology- and transcriptome-supervised cell-type heads retrainable from public data. Slides are read from cloud repositories and per-slide outputs are written to QuPath and OMERO. The same workflow is AI-agent callable through a Model Context Protocol endpoint. Applying it to TCGA cohorts recovered known immune and molecular associations.