GraphPatch

GraphPatch is a web app (with optional CLI) that continuously validates and repairs enterprise knowledge graphs. It monitors RDF/Property Graph datasets and their pipelines, detects schema drift, missing constraints, duplicate entities, and contradictory facts, then proposes “patches” (merge/split entities, add constraints, rewrite mappings, or quarantine bad sources). It connects to Neo4j, Amazon Neptune, and common ETL tools, and runs scheduled checks with alerting to Slack/Jira. The AI layer explains anomalies in plain language, suggests likely root causes (e.g., upstream API change), and generates safe, reviewable remediation plans rather than auto-editing production graphs. The MVP focuses on high-signal checks that teams actually care about: entity resolution regressions, ontology version mismatches, and pipeline-induced data corruption. It’s positioned as reliability tooling for knowledge graphs, not another graph database or generic data quality dashboard.