After a bunch of back and forth with myself I'm finally beginning to get a handle on a (probably not the best) way to create the "server" side of CaptureBot.
The ability to interact with the host system remotely was always a key feature, but I'd never quite figured out how I wanted to do it. I think creating a strong server/ client model will provide the most flexibility going forward.
The key realization for me was deciding I don't need to create a full REST API. I only need to provide access to a limited set of URIs that map to methods which in turn call to AppleScript handlers that communicate with Capture One on the host system.
Once this is all in place the client side can be anything from a full instance of CaptureBot to an iOS Shortcut to a curl command.
emory.coffee is one server in the network