Use of Computer Vision Activities with one single node Orchestrator application

So I have been researching for a bit on this topic but to me it’s unclear on how we are supposed to use the computer vision activities in our scenario:

To Quickly explain our scenario:

We have one single node enterprise Orchestrating application which is orchestrating a number of RPA processes across our robots.

We recently started adopting the computer vision activities. I am however having some mixed feelings in regard to the set-up:

From what I understand: to be able to use the Computer Vision activities we need to connect to a computer vision server.

To do this there are 2 options:

  1. Use the Community CV ML Server by specifying URL: and an API Key.
    –> The problem I see with this is that we are not able to retrieve an API key from our organisation(Orchestrator) in any way. This would mean that we are supposed to create a community user on the community cloud platform to work this way? --> Strange!

  2. Deploy our own ML Server:
    –> The problem I see with this is that we should provide the infrastructure to deploy the ML server to.
    Additionally, I fear we will be missing out on the community model teaching itself based on all of the incoming information.

One more question: Do the computer vision activities come at a licensing cost?

Thanks in advance for your responses!