Skip to main content
Volumes hold files your workloads share; artifacts hold downloadable task output. Both belong to the selected workspace. Repeatable workflows declare volumes and secret dependencies in Python, and LazyCloud creates a declared volume on first mount. These commands suit one-off work and inspection.

lazycloud volume

lazycloud disk

A deploy creates the disks its pods declare, so there is no create.

Files

The example below creates a volume before any workload uses it; once a workload has mounted its declared volume, volume create is unnecessary. Volume paths are lazycloud://<volume>/<path>. The scheme is optional except on the source of a download, where it marks the path as remote.
This uploads a local checkpoint, downloads a copy, and archives the remote file. uv run lazycloud rm models/archive/model.safetensors removes the archived file. Deleting a volume removes all its files.

lazycloud secret

A value typed on the command line lands in your shell history. Reading it from a variable or a file avoids that: uv run lazycloud secret create OPENAI_API_KEY "$OPENAI_API_KEY".

lazycloud artifact

artifact list finds files saved by tasks, and artifact usage reports their storage use:
uv run lazycloud artifact delete <artifact-id> removes a saved file. Lists return a next_cursor, and --cursor <cursor> fetches the next page. See Artifacts for downloads and retention.
To change your default workspace, run uv run lazycloud workspace use <name>. Commands with a --workspace option can override it for one call.