We are just getting started on our Medplum journey and bringing a new customer onboard in the coming months. We want medplum to run this client. Here is what we are thinking right now, customer doesn't currently have Fhir, they are able to provide an HL7 v2 stream. We are trying to get them to deploy the Agent, but getting fightback on that, so assuming I can't, I believe the best approach is to do the following: 1. Setup a VPN between the two locations 2. Deploy a Medplum Agent in our network 3. The Medplum agent will deploy HL7 to the server 4. We need to write bots that capture the data and either route it back to our app API, or store it, or both? For 4, is there a better approach others have used? For a bit of background, we are self-hosted in AWS. Our application has a fully operational API, integrates with SQS, so we can kinda do what we want to do from this side. I think our preference is to store the data in PSQL Fhir store at least temporarily.

#4 is the only way the Medplum processes messages from the Agent. HL7v2 processing tends to be business specific

no, I am asking as much because this is a very new world and its what I have pieced together from youtube videos, my limited experience with Medplum, Mirth and so on. it also helps me understand best approach based on what we are hearing from our customers.

That's right. "Bots" are basically Medplum plugins - you write code to process incoming data, write to the Medplum data store, and/or write to external systems