Agentic Spatial Pathologist#
Agentic workflows for spatial pathology.
Agentic Spatial Pathologist is an evidence-generating workbench for building, running, and reviewing pathology-aware Xenium analyses. The public Python package and CLI remain named spatho.
The platform narrative is: stGPT learns reusable contour/region morpho-molecular representations; spatho plans, validates, and turns them into auditable spatial pathology evidence.
The documentation focuses on practical deployment paths first, from local workstations to PDC/HPC environments with a self-hosted pathology AI backend.
Run the ATERA WTA breast PDC walkthrough and inspect the generated workflow artifacts.
Install and run spatho on a local workstation for development or lightweight analysis.
Deploy the local pathology AI service with vLLM, embeddings, reranking, and Qdrant.
Understand the repository layout, development workflow, and compatibility layer.
Review the roadmap and package release notes for the spatho distribution.
Plan the optional stGPT foundation-model runtime, region-first artifact contract, and agentic evidence loop.
Use stGPT region artifacts, scGPT-like RNA mapping, pathway activity, PLIP morphology, and lightweight niche fusion as auditable evidence.
Read design notes for Xenium RNA/protein alignment, commercialization, and packaging.