S3 Gateway

Buckets and keys

Bucket = connected provider

A bucket in Uploom's S3 surface corresponds to one of your connected cloud accounts. The bucket name is derived from the connection (the provider type and a per-connection identifier). You can see the exact mapping for your account by listing buckets from any S3 client.

Listing buckets

The standard ListBuckets call returns every connected account for the authenticated user:

aws s3 ls
# 2026-06-13 13:42:53 bucket-a
# 2026-06-13 13:42:53 bucket-b

Keys are paths

A key is the S3 path after the bucket. It's mapped to the provider's native folder structure, with / as the separator. Folders are conventional — S3 is a flat key/value store, and most cloud providers don't have a real folder type.

Folder markers

To create the appearance of a folder, the gateway supports a trailing-slash convention: PUT /<bucket>/folder/ with an empty body creates a folder marker.

aws s3 cp /dev/null s3://bucket/photos/2024/
# Creates a 0-byte object named "photos/2024/" — Uploom shows it as a folder

Listing keys

Use the standard S3 listing operations. The delimiter=/ query parameter is the easiest way to get a folder-by-folder view:

aws s3 ls s3://bucket/photos/ --recursive
# 2026-01-12 14:00:00      12345 photos/2024/cat.jpg
# 2026-02-20 09:15:00      89012 photos/2024/dog.jpg
# 2026-03-05 16:42:00        450 photos/cat-of-the-month.txt

Cross-bucket operations

CopyObject across buckets is not implemented. To move data between two of your connected providers, use the dashboard's "move to" action, which copies files between connections.