Test validation rules, data contracts and pipelines against the problems real data has: missing values, malformed emails, impossible dates, outliers and duplicate keys.
Free trial for every new account. No credit card.
What to look for in the screenshot above. These are the details that usually break in hand-made or random test data.
From 1% to 50% of rows; the rest stay perfectly consistent.
Nulls, formatting errors, outliers, impossible dates, broken totals and duplicates.
For SQL output, defects stay within what the database accepts, so the file still loads.
Use a seed to get exactly the same dataset, and the same defects, every time.
Describe your data in plain English, or upload your SQL schema for PostgreSQL, MySQL / MariaDB, SQL Server and SQLite.
Check the tables, columns and rules the AI proposes, and adjust anything you like.
Choose how many rows you need and download a ready-to-run SQL script, CSV, Excel (one sheet per table) or JSON.
Yes. Columns that other rows depend on, like primary and foreign keys, are protected unless you ask otherwise.
Yes. The defects are typical of real pipelines, which makes them good fixtures for data tests.
No. Anomalies are applied after the text is written, so the AI never writes about a damaged value.
Describe your tables or upload your schema. The free trial needs no credit card.
Start free