Docs

S3-compatible storage access

Use the AWS CLI, boto3, rclone, Cyberduck or any S3 SDK with your project's buckets: create an access key, point the client at your endpoint and use region auto.

Updated Oct 16, 2026Open in console

Every project's storage also speaks the S3 protocol. Objects written over S3 are the same objects you see in the console and the storage API: they count towards the same limits and show up in the audit log.

/storage/v1/s3SDK / storage APIuser session, bucket rulesS3 clientaccess key (server-side)Project gateway/storage/v1/…Your bucketsone copy of every object

Create an access key

  1. Open Storage → S3 connection

    Owners and admins can create keys. The page also shows your endpoint.

  2. Create a key and copy the secret

    The secret is shown once and stored encrypted. Put it in your secret manager.

  3. Configure your client

    Path-style addressing, the endpoint from the page, region auto.

Client setup

export AWS_ACCESS_KEY_ID=<access-key-id>
export AWS_SECRET_ACCESS_KEY=<secret>

aws s3 ls --endpoint-url https://<project-ref>.sovaria.in/storage/v1/s3 --region auto
aws s3 cp ./backup.tar.gz s3://backups/2026-10-16.tar.gz \
  --endpoint-url https://<project-ref>.sovaria.in/storage/v1/s3 --region auto

What is supported

  • Buckets: list, create, delete.
  • Objects: list (v1 and v2, prefixes and delimiters), get (with ranges), head, put, copy, delete, batch delete.
  • Multipart uploads and presigned URLs.

S3 access is included on every plan; storage size, file size and egress follow your plan — see Plan limits.

Something unclear or missing on this page?Tell us