Data Anomaly Finder

Upload or paste CSV, TSV, JSON or JSON-lines data and get a list of values that look wrong: numeric outliers, duplicate rows and IDs, dates in a different format, mixed types, rare category values, missing-value spikes and end-before-start dates. Every finding comes with a plain-language reason. These are statistical rules (median/MAD, IQR, frequency and shape checks), not a trained model, so treat the results as leads to review. Nothing is stored.