← All guides

How-to

How to create an S3 bucket without touching the AWS console

Drag an S3 node onto a canvas, set a few fields, and provision a real bucket in your AWS account — a visual walkthrough of creating S3 buckets with Korve.

Updated 2026-07-18

The AWS console works, but it's a lot of tabs for one bucket. Here's how to create a real S3 bucket by dragging a node onto a canvas instead — with versioning, encryption, and access control set the way you want, and no CloudFormation to write.

1. Drop an S3 node

With AWS connected, open a workflow and drag S3 from the service sidebar onto the canvas. A node appears. Nothing has been created in AWS yet — the canvas is a document you're editing, and resources are only provisioned when you deploy.

2. Configure the bucket

Select the node and its configuration panel opens. The fields map one-to-one to what S3 actually offers:

• Bucket name — validated against S3's DNS-compliant naming rules as you type, so you catch an invalid name before AWS rejects it.
• Region — where the bucket lives.
• Access control — private (the safe default), public-read, or public-read-write.
• Versioning — keep old versions of overwritten objects.
• Encryption — server-side encryption, with a KMS key selector if you want your own key.

Every field you'd hunt through three console screens for is on one panel.

3. Validate, then deploy

Hit Deploy and Korve shows a review first. Run a dry run (free, changes nothing) to confirm exactly what will be created. When it looks right, deploy for real — Korve calls the S3 API with your settings and the bucket appears in your account, region, and console within seconds. The node shows a live sync status so you know it's real, not a diagram.

4. Work with the bucket

Once it exists, the same node lets you browse objects, upload and download files, edit the bucket policy as JSON, and set lifecycle and CORS rules — the full surface, without leaving the canvas.

Why do it visually

A bucket alone is easy anywhere. The point of the canvas shows up the moment the bucket connects to something else — a CloudFront distribution in front of it, a Lambda that writes to it, an EC2 instance that reads from it. Draw the connection and Korve creates the IAM permissions that wiring implies. See wiring EC2 to S3 for exactly what that means.

Tearing it down

Made it to try it out? Deleting is just as direct — remove the node and deploy, or delete it from the resource list. Nothing lingers billing you.