S3 buckets
A bucket is added like any other connection and browsed with the same file explorer as an SFTP server: you go in, see what is there, upload and download.
What you need to enter
Picking "S3" as the connection type changes the form to ask for what matters:
- Service endpoint: s3.amazonaws.com for AWS, or your compatible service address.
- Bucket: the exact bucket name.
- Region: us-east-1 if you are not sure, which most compatible services ignore anyway.
- URL style: subdomain for AWS, path for MinIO, Versity and almost anything self-hosted.
- Access key and secret key.
Path or subdomain
This is the one option that usually trips people up. AWS builds URLs as bucket.endpoint/file (subdomain); self-hosted compatible services almost always need endpoint/bucket/file (path).
Getting it wrong does not produce a clear error: it produces a 404 or a 403 that looks like a permissions problem. That is why it is asked outright instead of guessed.
Saved only if it works
On save, SHTERA asks for a one-object listing with the keys you entered. That validates the endpoint, bucket, region, URL style and both keys in one go. If anything is wrong, you are told and nothing is saved.
What a bucket looks like
S3 has no folders: it has names with slashes in them. The explorer groups them and shows them as folders, which is what the AWS console does too.
Three things therefore behave differently from a disk: an empty folder exists only as a marker, renaming is copy-then-delete (for a folder, everything inside it), and deleting a folder deletes all of its objects.
The secret key
It is stored encrypted in the folder, like any other server password: the SHTERA server never sees it in the clear. The endpoint, bucket and access key are stored as data, because they are not secrets.