Cloudflare D1: a zero-infrastructure database for SaaS on Workers

Cloudflare D1 is a serverless SQLite database that lives next to your Worker. It's the default database when you deploy ShipAny TanStack to Cloudflare: no servers, no connection strings, no pooling to think about.

What ShipAny handles

• A D1 driver that reads the DB binding from the Workers environment — no URL needed.
• The SQLite schema for all built-in tables, applied with Drizzle migrations.
• Bundle trimming: when DATABASE_PROVIDER is d1, the other database drivers are stubbed out of the Worker bundle.
• One-command setup with the deploy-cloudflare agent skill: it creates the database, writes the binding, applies migrations, seeds roles and permissions, and is safe to re-run.

Setup

1. Create the database: npx wrangler d1 create <name> and copy the database_id.
2. In wrangler.jsonc, set vars.DATABASE_PROVIDER to "d1" and add the binding:

"d1_databases": [
  { "binding": "DB", "database_name": "<name>", "database_id": "<id>", "migrations_dir": "drizzle" }
]

3. Generate and apply migrations: pnpm db:generate, then npx wrangler d1 migrations apply <name> --remote.
4. Seed roles and permissions, then deploy with pnpm run cf:deploy.

Tips

• D1 is for Cloudflare Workers deploys. Locally, the same schema runs on a plain SQLite file.
• migrations_dir must point at drizzle, or wrangler won't find any migrations to apply.
• Outgrowing D1 or need Postgres features? Switch to PostgreSQL through Hyperdrive.