Agent versions

Every save of an agent creates a snapshot. The History tab on the agent surfaces the snapshots in reverse chronological order; comparing two snapshots shows the diff of what changed, and restoring a past snapshot replaces the current state with that version. There is no manual save versus auto-save distinction — every persisted change is a version.

The mechanic is small but load-bearing. Most teams adjust an agent's instructions weekly; without the history, the team would never trust the edits.

A worked diff

Open the agent and switch to History. The list shows Current version at the top and every prior Snapshot version below, with the author and timestamp on each row. Click two rows and Compare changes opens a side-by-side diff; the changed fields highlight. Close the diff to return to the list.

Restoring a version

Open any past snapshot and click Restore this version. The agent's current state is overwritten with the snapshot, and the act of restoring itself creates a new snapshot on the timeline — restores are not destructive, just additive. Chats already running against the previous current version continue to run on it until they end; the restored version applies to new chats from that point on.

What gets versioned

Versioning covers instructions, model picks, knowledge bindings, tool toggles, conversation starters, and metadata. It does not cover the underlying knowledge sources themselves — replacing a document the agent is bound to changes what the agent retrieves without bumping the agent's version. To audit a knowledge change, see Audit logs.

Where this fits

Versions is the agent's safety net for the same reason git is the codebase's: anything saved is recoverable. The companion page is Audit logs — it covers the org-wide who-did-what trail; versions cover the per-agent what-was-it trail.