
Cluster
Cluster is the inventory of everything that runs: every agent, service and app, with its owner, its configuration and its health. It is where you see what each agent may write, which internal recipients it may reach, and which version of its rules is loaded.
At a glance
- Registration and manifests for everything that runs
- Health collection, incidents and analytics
- Git watch and architecture scan across every repository
In our own company. Our own Cluster tracks every agent, service and app we run.

Cluster on fictional Example Co demo data. This screen labels the levels in its own short form (L0 Observe only or Report only, L1 Suggest, L2 Acts reversibly); the ladder on this site is L0 Observe, L1 Draft, never send, L2 Act reversibly, L3 Decide.
Know everything that runs.
One inventory
Agents, services and apps are registered with a manifest, so nothing runs unlisted.
Each agent's limits
Rules, model, grants, its internal recipient list and its message limits for one agent on one page.
Loaded against latest
See whether the configuration an agent is running is the latest one that was published.
Health and incidents
Health collection, incidents and analytics across the whole estate.
Code watch
Git watch and an architecture scan across every repository.
History
Configuration changes are kept in a history you can read back.
See Cluster on your own work.
Request a 45-minute call. We show it running in our company and what it would look like in yours.

