Snapshots

A snapshot is a reusable checkpoint of a sandbox. It preserves the sandbox’s filesystem and runtime state so that you can later start a new sandbox from the same point instead of rebuilding the environment and rerunning setup work.

• Templates are stored as snapshots. A template build commits one snapshot; the template ID is an alias that resolves to it.
• Warm-started sandboxes restore the state of a committed template or snapshot.
• Running sandboxes can produce new snapshots, capturing their current state for later reuse or branching.

What a Snapshot Preserves

A snapshot preserves:

• VM state and memory, allowing running processes to continue from the captured point.
• The root filesystem and its changes.
• Attached-drive contents and attachment metadata.
• Mounted-volume contents, mount paths, sizes, and access modes. Starting from the snapshot creates new volumes from the captured contents rather than reusing the source volume identities.
• CPU, memory, and disk settings, together with command context such as environment variables, working directory, user, and startup commands.

Where to Go Next:

• Create a Snapshot — capture a reusable checkpoint from a running sandbox.
• Manage Snapshots — list, inspect, and delete snapshots.
• Rootfs as an OCI Image — publish or export only the captured root filesystem.
• Optional P2P Visibility — accelerate committed snapshot artifact distribution between nodes.