Deploy to Cloudflare

By the end of this tutorial your notes API (HTTP routes, D1 database, and GraphQL endpoint) will be live on Cloudflare's edge network.

This is the first step that needs a Cloudflare account (the free tier is fine).

Get credentials

The base CLI drives Cloudflare through two environment variables:

• CLOUDFLARE_ACCOUNT_ID: on your Cloudflare dashboard's overview page.
• CLOUDFLARE_API_TOKEN: create one at My Profile → API Tokens with the Edit Cloudflare Workers template (add D1 edit permissions for the database steps).

Put them in your worker's env.toml — it's gitignored precisely so secrets like these never reach your repo:

[Production]
CLOUDFLARE_ACCOUNT_ID = "your-account-id"
CLOUDFLARE_API_TOKEN = "your-api-token"

The CLI loads the section matching the environment you target.

Create the production database

Locally, wrangler simulated D1 with a placeholder id. Production needs the real thing:

npx wrangler d1 create app-db

Copy the database_id it prints into the production environment of wrangler.toml:

[env.Production]
name = "app"
d1_databases = [
    { binding = 'DATABASE', database_name = 'app-db', database_id = 'your-real-database-id', migrations_table = '__drizzle_migrations_app', migrations_dir = './database/@default/drizzle/migrations' },
]

Each environment block is a complete description of that environment — same binding name, different database.

Release your migrations

Base gates deploys on deliberate schema changes: a migration must be released before a worker that depends on it can ship.

npx base orm migration:release -w app

This promotes your current migrations to the released list (a file in your repo — commit it). Then apply them to the production database:

npx base orm migration:run -w app -e Production

Check, then deploy

Commit your work first — production deploys require a clean git tree, so what ships is exactly what's in history. Then let the CLI validate everything (settings, wrangler config, ORM bindings) before it flies:

npx base check app
npx base deploy -w app --environment Production

deploy bundles your worker with esbuild and publishes it via wrangler. The environment flag is required on deploys — there is no accidental default target.

When it finishes, your worker is live on your workers.dev subdomain:

curl https://app.<your-subdomain>.workers.dev/notes

Create a note against production, query it back over /graphql — everything you built in the last four tutorials is now running at the edge.

Watch it run

Stream live logs from the deployed worker whenever you need to see what's happening:

npx base tail -w app -e Production

And note what the scaffold already gave you: the .github/workflows/ directory contains CI and deploy workflows, so wiring this same deploy into GitHub Actions is configuration, not new work.

Where to go next

You've gone from empty directory to a typed, validated, database-backed API with REST and GraphQL, deployed globally. One tutorial remains, and it's the payoff: Call Your Worker with RPC turns your API into typed method calls. Beyond that:

• How-to guides: focused recipes for routing, ORM relations and filters, GraphQL patterns, dependency injection, validation, access control, queues, and scheduled tasks.
• Reference: every class, decorator, and type, generated from the source code itself.