Sketch the VM, keep the .tf

Most Terraform in a company is the same resource written for the twentieth time: a VM, its NIC, its security group, the variables for the region. The interesting part is not typing it. The interesting part is whether the subnet it points at exists, whether it has egress, and whether the credential that will run it is allowed to.

Control Cloud lets you drop the VM on the map of the cloud you already have, checks it against the live state the moment it lands, and then writes the Terraform for it.

Four files, regular HCL

Compute, network security group, providers, variables. Regular HCL, the kind you would have written by hand, with the names you typed on the canvas. Copy All or download, and it goes into your own repo like any other module.

Checked before it is code

The VM is rejected on the canvas if it is not connected to any network. That happens before a single line of Terraform exists, so the file you take away already points at a subnet that is there. A green terraform plan on a VM with the wrong subnet is still a red apply; the canvas catches that one step earlier.

Apply where you decide

Run it in your CI with your own key, or let Control Cloud apply it with the credential you scoped for that. Either way the code works with or without us: no wrapper, no custom provider, no lock-in.

Watch the 25-second clip · Full 60-second demo

Try it

Free plan: 1 operator, 1 provider, no card. Read-only API access to map, write access only where you grant it.

Which resource do you still write by hand most often? I answer every comment.