@byronwade/seed

Schema-driven fake data generation for Fakebase. Generate realistic, referentially-correct, deterministic rows straight from your schema — no hand-written seed files.

pnpm add @byronwade/seed

import { seedClient } from "@byronwade/seed";

// Idempotent — skips tables that already have rows.
await seedClient(supabase, schema, { rowsPerTable: 20 });

• Foreign-key correct — children reference real parent rows; inserts run in FK order.
• Deterministic — same seed ⇒ identical data (stable reloads, clean diffs).
• Convention-first — column type + name inference (email, avatar_url, price…), with per-column overrides when you need control.
• Honest — describeResolution() shows how every column resolves.
• Pluggable — zero-dependency built-in provider; install @byronwade/seed-faker for @faker-js/faker data and locales.

Exports: seedClient, generateRows, describeResolution, createBuiltinProvider, orderTables, createRng, and the DataProvider interface.

Full guide: docs/seeding.md.

Dev/prototype tool — not production infrastructure.