# Buckets

Source: https://docs.quake.ai/docs/object/concepts/buckets
Markdown: https://docs.quake.ai/docs/object/concepts/buckets.md

---

# Buckets

A **bucket** is the top-level container in object storage: a named space that holds objects (files) addressed by keys. On Quake AI buckets live in **[object storage](/docs/object/concepts/object-storage)** (OpenStack Swift) and appear in the Console under object storage.

S3 clients call the same object a bucket. Swift documentation sometimes says **container**; customer-facing copy on Quake AI standardizes on **bucket**.

## On Quake AI

- Buckets are project-scoped.
- Access uses **[S3 credentials](/docs/tools/s3-credentials)** or Swift/Keystone patterns depending on the client.
- URLs follow `https://object.<region>.rumble.cloud/<bucket>/<key>` for path-style access; virtual-hosted style is also supported for S3 clients when configured.

Bucket policies, versioning, and CORS are configured per bucket. See object how-tos for procedures.

## Related vocabulary

| Term | Same thing? |
|---|---|
| S3 bucket | Yes |
| GCS bucket | Yes |
| Azure blob container | Yes (blob container ≈ bucket) |
| Swift container | Yes (internal name) |
| Block volume | No (see [volumes](/docs/block/concepts/volumes)) |

For S3-specific client setup, see [S3 on Quake AI](/docs/object/concepts/s3).

## What to read next

- [Object storage](/docs/object/concepts/object-storage): architecture and access patterns
- [Create a container](/docs/object/how-to/create-container): first bucket in the Console
- [S3 on Quake AI](/docs/object/concepts/s3): SDK endpoints and compatibility notes