# S3 on Quake AI Source: https://docs.quake.ai/docs/object/concepts/s3 Markdown: https://docs.quake.ai/docs/object/concepts/s3.md --- # S3 on Quake AI **Amazon S3** naming is everywhere in tooling, SDKs, and migration guides. On Quake AI, S3-compatible access targets **[object storage](/docs/object/concepts/object-storage)** backed by OpenStack Swift with an S3 API gateway. You create **buckets** (Swift containers), upload **objects**, and authenticate with **[S3 credentials](/docs/tools/s3-credentials)** or Keystone-derived keys. The product documentation uses **object storage** and **bucket** as primary nouns; **S3** describes the API dialect your client speaks. ## On Quake AI - **Endpoint pattern:** `https://object..rumble.cloud` (see [service endpoints](/docs/tools/service-endpoints)) - **Console label:** Object storage buckets under the storage section - **CLI/SDK:** AWS CLI, `boto3`, `rclone`, and similar tools work when you set the custom endpoint and region - **Auth:** S3 access keys from the Console; not the same as OpenStack application credentials S3 compatibility covers the operations most applications need; some AWS-only features (certain lifecycle rules, event notifications, IAM bucket policies) differ or are absent. Validate critical calls against [S3 compatibility](/docs/object/faq) and the object API reference before cutover. ## What other providers call this | Provider | Their term | On Quake AI | |---|---|---| | AWS | S3 bucket | Object storage bucket | | Google Cloud | Cloud Storage bucket | Object storage bucket | | Azure | Blob container | Object storage bucket | | DigitalOcean | Space | Object storage bucket | | Hetzner | Object Storage bucket | Object storage bucket | ## What to read next - [Object storage](/docs/object/concepts/object-storage): architecture, access patterns, and Swift relationship - [Create S3 credentials](/docs/object/how-to/create-s3-credentials): keys for SDK and CLI access - [Migrate from Amazon S3](/docs/object/migration/migrate-from-s3): rclone and endpoint cutover