Generate believable fake data. Without piping your schema into a stranger's server.
Describe your fields — names, emails, dates, UUIDs, numbers, enums —
pick a row count, and get back ready-to-paste
JSON, CSV or SQL inserts. It is
seeded, so the same seed always produces the same rows. Everything
happens in your browser; the schema is never uploaded.
FAQ
Is my data sent anywhere?+
No. Your field definitions and every generated row are produced entirely in your browser — nothing is uploaded, and there is no schema to leak. We count anonymous, aggregate usage on our own server (a page view, that a batch was generated) — never your schema, your field names, or the data we generate for you.
Why is there a seed, and what does it do?+
The generator is built on a seeded pseudo-random number generator, so the same seed plus the same schema always produces the exact same rows. That makes your fixtures reproducible — commit the seed and a teammate (or a CI run) regenerates byte-identical test data. Change the seed for a fresh batch; the seed box shows the current one so you can pin it.
Is the fake data realistic enough to seed a database?+
Yes — names, emails, dates, UUIDs and enums are drawn from small curated word lists, and the CSV and SQL exports are properly escaped (commas and quotes are handled, single quotes in SQL are doubled) so they paste straight into a table. It is intentionally a curated subset, not a giant dictionary; relational/foreign-key generation, locale packs and multi-thousand-row exports are the paid tier. No login, no payment — this one is free.