What problem does it solve? Populating development, demo, or test databases with realistic sample data is tedious and error-prone: developers must respect foreign key ordering, handle platform limits like D1 batch constraints, and avoid generic placeholder data that fails to exercise real application behavior. ## Core Features & Use Cases - Schema-aware generation: Reads Drizzle schemas, SQL migrations, or Prisma files to build a dependency graph and insert parent tables before children. - Realistic domain data: Produces names, emails, dates, and prices matched to the application domain instead of "test123" placeholders, with optional deterministic seeded randomness. - Idempotent output: Emits TypeScript (Drizzle) or raw SQL seed files that are safe to re-run via delete-then-insert or upsert strategies, with D1-specific batching handled. - Use Case: You just defined a Drizzle schema for a blog with users, posts, and comments. Ask for a seed script and receive a ready-to-run scripts/seed.ts that inserts 20 users, 50 posts, and comments in correct foreign-key order, batched for D1 limits. ## Quick Start Ask the agent to generate a seed script for your database by pointing it at your Drizzle schema or SQL migrations and specifying the purpose (dev, demo, or testing) and row volume.
