PLATFORM
Ten capabilities. One job model.
Every one of these runs the same way: a request becomes a queued job, the job streams its output live, and what it did lands in the audit ledger attributed to whoever asked for it. The difference between managing a Frappe site and an EKS cluster is the adapter at the end, not the model in front of it.
INFRASTRUCTURE
The machines, wherever they run.APPLICATIONS
What runs on top of them.DATA
The part you cannot recreate.GOVERNANCE
Who may do what, and what happened.AI ENGINE
BETAAn engineer that reads your fleet, and asks before it acts.
It sits across all ten capabilities above, with your own provider key, and it cannot act directly — it writes a proposal, a human reviews the diff, and only then does the work enter the same queue everything else runs through.
explore →One control plane, whichever layer you are on.
Install it on a host you already have and connect the first thing read-only.