SchemaSleuth

SchemaSleuth is a web app (with optional desktop agent) that helps teams build and maintain a practical knowledge graph from messy, real-world sources: spreadsheets, Postgres tables, CSV exports, and internal docs. It auto-suggests entities, relationships, and constraints, then lets users approve changes with a lightweight review workflow. The core value is not “a perfect ontology,” but a graph that stays accurate as data changes. It continuously monitors upstream sources, flags schema drift (renamed columns, new codes, broken joins), and proposes safe migrations to keep the graph consistent. It also generates human-readable documentation and a change log so non-experts can understand what the graph means. This is an AI app + traditional app combo: AI accelerates mapping and drift detection, while deterministic rules enforce consistency and prevent hallucinated structure.