AI agent deployment

Don’t want to memorize Wrangler commands? Copy-paste the prompt below into your AI agent (Cursor, Claude, etc.) and it will handle everything.

First deploy

Deploy my app to Cloudflare Workers.

Steps:
1. Run `bun run build` to build client assets
2. Run `wrangler d1 migrations apply <database-name> --remote` to apply database migrations
   (read the database name from wrangler.toml)
3. Run `wrangler deploy` to deploy the Worker
4. Curl the health endpoint to verify it's live
5. Tell me the production URL

Set secrets

Set these production secrets for my app on Cloudflare Workers:
- NODE_ENV = production
- RESEND_API_KEY (I'll provide the value)
- GOOGLE_CLIENT_ID (I'll provide the value)
- GOOGLE_CLIENT_SECRET (I'll provide the value)

Use `wrangler secret put` for each. Ask me for the values before running the commands.

Update an existing deployment

Deploy the latest code to my app on Cloudflare Workers.
Build, apply any new migrations, then deploy. Verify it's live after.

Download production data to local

Export my production D1 database to a local SQL file and import it into my local D1.
Read the database name from wrangler.toml, then use
`wrangler d1 execute <database-name> --remote --export ./backup.sql` and
`wrangler d1 execute <database-name> --local --file ./backup.sql`.

That’s it. Your AI agent reads these prompts, runs the commands, and tells you the result. You never touch Wrangler directly.