Seeding

This page stays deliberately general. The seed files are the source of truth.

What seeding does

• Creates a repeatable local dataset for development and tests.
• Uses a namespace so multiple contributors can share a database without colliding.
• Seeds users first, then orgs, then follow-up data that depends on those orgs.

How it works

• pnpm seed validates that DATABASE_URL points to a local or explicitly allowed dev database.
• The current namespace is cleared before reseeding, so reruns stay predictable.
• The namespace comes from SEED_NAMESPACE, or falls back to your git user name, system user name, or a generated run id.
• Seeded emails and display names are namespaced so contributor data stays isolated.

What gets seeded

• Seed users for sign-in and test flows.
• Demo / sample org data.
• Supporting data that depends on the seeded orgs, such as conversion data and invite fixtures.

Demo sessions

• The demo button uses a separate runtime seed flow to build one isolated demo org per visitor.
• That demo org is pre-seeded with sample data, then reset automatically after use.
• Each demo launch also writes a persistent analytics record so the admin area can track demo usage without depending on rows that later get cleaned up.
• This is separate from pnpm seed, which builds the shared local/test dataset.

Cleanup

• Use pnpm seed:clean when you only want to remove your own namespaced seed data.
• It leaves other contributors' data untouched.

Notes

• If you need the exact model layout, see Data Models.
• If you need the enum values used by the seed data, see Enums.