wrangler

Run wrangler deploy from your project directory after configuring wrangler.jsonc with your Worker name, main entry point, and compatibility_date. Use wrangler deploy --dry-run first to validate changes without deploying, and wrangler deploy --env staging to target a specific environment.

Use wrangler secret put API_KEY for an interactive prompt, which is the preferred secure method. Never pass secret values as command arguments or echo them; for bulk operations use wrangler secret bulk with a JSON file that is not committed to version control.

Prefer wrangler.jsonc because newer Wrangler features are JSON-only. Add the $schema field pointing to node_modules/wrangler/config-schema.json for autocompletion and validation of config fields and binding shapes.

By default wrangler dev uses local storage simulation for bindings. Set remote: true on specific bindings in your config to connect to real resources, which is required for Workers AI and recommended for Vectorize and Browser Rendering.

The binding name in your Worker code must exactly match the binding name in wrangler.jsonc. After changing config, run wrangler types to regenerate TypeScript definitions and verify the binding names align.

Create migrations with wrangler d1 migrations create, then apply them locally using wrangler d1 migrations apply my-database --local or to production with --remote. List pending migrations with wrangler d1 migrations list before applying.