Every provider, one way of working.
Many teams end up on more than one provider: rented metal here, a VPS there, something still in a hyperscaler, and a rack of their own. Each has its own console, its own API, and its own way of doing things, and none of it carries over.
The same workflow on every substrate
One dashboard, one API, and one Terraform provider cover your own rack, rented metal, and any VPS. No console per vendor, and no separate automation path for each one.
The same image boots anywhere
Servers are versioned images written straight to disk, so the build you tested runs identically wherever it lands. Moving a workload to another provider is a reprovision, not a redesign.
Every provider lands on the same private network
Machines join one encrypted mesh regardless of who they are rented from, connecting through NAT and firewalls with nothing to open. That flat network is what lets a cluster span vendors at all.