Connect to object storage with S3 access keys

Object storage works with tools that speak Amazon S3: rclone, the AWS CLI, Cyberduck, backup software and the AWS SDKs.

Your endpoint and region

Open Object Storage. The Endpoint and Region to use are shown there, with ready-made settings for rclone, the AWS CLI and Python (boto3) that already have them filled in. Copy those rather than typing them from memory.

Path-style addressing

Use path-style addressing, where the bucket name goes in the path (endpoint/bucket/file), not in the hostname (bucket.endpoint/file). Most tools need telling:

• rclone: force_path_style = true
• boto3: config=botocore.config.Config(s3={"addressing_style": "path"})
• AWS CLI: run aws configure set default.s3.addressing_style path once, and pass --endpoint-url with the endpoint on every command

The rclone and boto3 snippets on the page already include this.

Create an access key

1. On Object Storage, under Access keys, choose Create an access key.
2. You are shown the access key and the secret key. Copy the secret now: we only ever show it once. If you lose it, revoke the key and make another.
3. Put both into your tool's settings.

Revoking a key stops anything using it immediately. Keys only work while your object storage plan is active.

Large files

A single upload request can be up to 100 MB. The AWS CLI, rclone and the SDKs send larger files in parts automatically.

Swift

If the page shows a Swift Storage URL instead of an S3 endpoint, S3 is not switched on for your account yet. Your buckets will be reachable over S3 as soon as it is, with nothing to move. Until then, ask us for a Swift token.

Last updated 21 October 2026