← All terms

Agent Archival

Moving an inactive AI agent into long-term storage while preserving its state for future restoration.

Agent Archival

Agent archival is the process of moving an agent out of active rotation and into a long-term, inactive state while preserving its underlying data so it can be restored later if needed. It is distinct from deletion: an archived agent's configuration, history, and stored state remain intact, but the agent is not running and is not expected to resume automatically.

How it works

Archiving an agent is typically an explicit action taken by an operator, rather than an automatic response to inactivity. Once archived, the agent no longer consumes active compute resources such as a running container, but its records, task history, and configuration are kept in storage. A separate restore action reverses the transition, bringing the agent back into a state where it can be resumed and assigned new tasks.

Archival vs auto-pause

Archival and idle auto-pause both remove an agent from active compute usage, but they serve different purposes. Auto-pause is a short-lived, automatic response designed to be reversed the moment new work arrives, and the agent wakes without operator intervention. Archival is a deliberate, longer-term decision, generally made when an agent has finished its purpose or is not expected to be used again soon, and it requires an explicit restore action to bring the agent back rather than happening automatically on task submission.

Why it matters

Over the life of a project, many agents are created for a specific purpose and then outlive their usefulness without being candidates for outright deletion; their history may be worth keeping for auditing, reference, or the possibility of future reuse. Archival gives operators a middle ground between an agent that is actively costing compute and one that is permanently gone, which matters when running a fleet of agents that grows over time and where storage is cheaper than compute.

Agent archival in Agenhood

Agenhood supports archiving an agent and restoring it later, alongside the separate pause and resume actions used for short-term inactivity. Archived agents can be restored through the web console or the REST API, returning them to a state where they can be resumed and given new tasks.

Get started

Deploy your fleet.

Put a fleet of sandboxed agents to work on your own infrastructure, provisioned in seconds and watched live from one console.

Get started

Admin-provisioned · Self-host in one command · Your data never leaves your VM