Docs

Bucket browser and object operations

The browser is designed to behave like a serious storage client and admin workspace, not a flat object dump.

WindowsmacOSLinux
BrowserPreviewURLs

Bucket administration, prefix navigation, search, folder semantics, metadata editing, and object actions.

  • Bucket administration controls
  • Prefix and folder behavior
  • Global search and local filters
  • Folder actions
  • Keyboard, metadata, and bulk operations
Applies to
WindowsmacOSLinux
Covers
BrowserPreviewURLs

Bucket browser and object operations

Bucket administration belongs in the same workflow

R2 Desk Pro treats bucket administration as part of day-to-day operations, not as a separate dashboard-only job.

Bucket-level actions include:

  • list buckets
  • create buckets
  • delete buckets
  • create buckets with location hints
  • choose Cloudflare-managed storage class where supported
  • work with jurisdiction-aware bucket settings

Bucket settings and administration

From the bucket settings workflow you can inspect or change:

  • bucket location
  • location hint
  • Cloudflare bucket storage class
  • jurisdiction
  • CORS rules
  • supported lifecycle rules
  • public r2.dev access
  • custom domains
  • event notifications for Cloudflare Queues

Treat those as operational controls, not edge features.

Understand the folder model

Cloudflare R2 uses object keys and prefixes. "Folders" in the UI are prefix groupings, not native filesystem directories.

That matters because:

  • Opening a folder navigates into a prefix.
  • Downloading a folder expands all objects under that prefix.
  • Deleting a folder deletes the objects under that prefix.
  • Moving or renaming a folder is a recursive object operation, not a metadata toggle.

The browser includes multiple navigation paths so users do not get trapped in long key paths:

  • Breadcrumbs for current prefix
  • Root and Up actions
  • Saved prefix shortcuts
  • Folder rows in the table

Use saved shortcuts for frequently used build output, media, or deployment prefixes.

Search and filtering

The UI exposes two different concepts:

Search entire bucket

Use this when the object may not be visible in the current page or prefix. This is backend-driven and intended for bucket-wide lookup.

Filter loaded rows

Use this when you already loaded the relevant prefix and want to narrow the current result set quickly.

Type and sort

Use filters and sort when navigating large prefixes:

  • Type: all, folders, files, images, documents
  • Sort: name, modified, size
  • Pagination keeps large result sets usable without pretending the bucket is locally indexed all at once.

File and folder actions

Available file actions include:

  • Details
  • Download
  • Rename
  • Move
  • Copy
  • Copy key
  • Copy public URL
  • Generate signed URL
  • Purge cache
  • Edit metadata
  • Delete

Available folder actions include:

  • Open folder
  • Download folder
  • Copy prefix
  • Delete contents

Folder download and delete expand recursively before the transfer or delete operation is queued.

Bulk selection

Bulk actions work for both files and folders.

When folders are selected:

  • Download selected resolves all descendant objects and writes them relative to the selected destination.
  • Delete selected resolves all descendant objects before issuing the delete batch.

This is intentional. A folder in R2 is not a single deletable object.

Object operations

Object workflows include:

  • rename
  • move
  • copy
  • delete
  • metadata editing
  • public URL copy
  • signed URL generation

These actions stay attached to the selected object so operators do not lose delivery context while making storage changes.

Keyboard shortcuts

The object table supports practical keyboard navigation:

  • Arrow Up / Arrow Down: move focus
  • Home / End: jump to start or end
  • Enter: open folder or details
  • Space: toggle selection
  • Delete: delete selected object
  • Ctrl/Cmd + C: copy object key

Object details

The details panel is where you verify metadata and link behavior before acting.

Typical uses:

  • Confirm content type
  • Check cache-control
  • Inspect ETag and last modified
  • Copy public URL or signed URL
  • Preview eligible image and text assets

Metadata editing fallback

If metadata rewrite is unsupported by the endpoint for a specific object path, the product treats those fields as read-only instead of pretending the write succeeded.

That fallback is deliberate. The app should report the endpoint boundary directly rather than implying that a metadata rewrite worked when it did not.