MEGA Hub
New
Docs/Get started

Xet Storage Overview#

Xet makes large files practical in versioned MEGA model, dataset, and Space repositories. It works with the existing repository revision model: commits, branches, tags, visibility, and repository permissions remain the way you organize and share a release.

When to use it#

Use the normal repository workflow for model weights, dataset shards, media, and other release artifacts that should be tied to a commit or tag. The MEGA CLI provides resumable upload and download commands for large trees.

BASH
mega upload-large-folder OWNER/REPOSITORY ./release --revision mainmega snapshot OWNER/REPOSITORY --revision v1.0 --local-dir ./release

Use Storage Buckets instead for mutable files that must change in place or be synchronized repeatedly.

What to expect#

  • Large transfers can resume after an interruption when you retry the same workflow and destination.
  • Repeated or incrementally updated content can reuse existing data, reducing unnecessary transfer work.
  • Repository access rules and visibility apply to large artifacts just as they do to other repository files.
  • A repository revision remains the unit you review, publish, and reproduce.

Keep releases reproducible#

Publish the model configuration, data description, and other required files in the same repository revision as the large artifacts. Pin consumers to a commit or tag, not a moving branch. See Hub Repositories and Large Files for release guidance.