Skip to main content
To recover remote worktree execution, keep the machine and execution IDs printed by the CLI. If a connection ends before you receive the result, inspect the existing execution before starting another.

​
Recover output

status does not wake the machine. If it is paused, run bead resume "$MACHINE_ID" and wait until it is running before recovering output; retrieve it before bead worktree up or another managed command retires the previous spool.
Output recovery does not replay the command. The spool retains up to 64 MiB and lasts until the next managed command or machine deletion. Save the output you need before either event.

​
Stop disconnected work

Use this for a service or command left running after its CLI disconnected. Managed commands report busy while a dev invocation owns the machine. Direct same-tenant command or file access can bypass worktree ownership; the next managed sync checks drift.

​
Recover a machine binding

If creation has an unknown outcome beyond its replay window, inspect existing resources and associate the matching machine. The template and shape must match your configuration.

​
Delete and prune

down retains the binding identity until destruction is acknowledged. It leaves templates and prepared snapshots retained separately. prune previews this installation’s confirmed removed-worktree orphans. --apply revalidates them before requesting deletion. Unknown or unmounted repositories are ineligible. Only paused orphan machines with matching binding metadata are eligible; running machines are skipped. Check machine lifetime for funded deadlines and pause behavior.