LineageLens

LineageLens is a web app (with optional CLI agent) that auto-builds end-to-end data lineage by parsing what already exists: Airflow/Prefect/Dagster metadata, dbt artifacts, warehouse query logs, and BI query histories. Instead of asking teams to manually document pipelines, it reconstructs table/column-level dependencies, freshness, and blast radius from real execution evidence. You get a searchable graph, impact analysis for schema changes, and a “what broke and why” view that ties incidents to upstream jobs and recent deployments. It’s a combination traditional + AI app: traditional ingestion/graph for correctness, plus AI to summarize incidents, generate change-risk reports, and suggest owners/runbooks. Reality check: this only works well if you can access logs/metadata; if your stack is opaque or heavily custom, results will be partial. Still, partial lineage is often better than none.