Running your model
A model that compiles is not the same as a model that behaves. One click starts a real OPC UA server exposing yours, so you can browse it with a real client before writing any server code.
Start it
Try Server in the editor toolbar. You get back an endpoint:
opc.tcp://<host>:<port>/opcua-server/<slug>
Point any OPC UA client at it — UaExpert,
node-opcua's CLI, or your own. Anonymous, no security policy: it is a sandbox
for looking at an address space, not a production endpoint.
What to look for
This is where modelling mistakes become obvious in a way the diagram does not always make plain:
- Is the hierarchy the shape you expected when you browse it?
- Did the members you marked mandatory actually materialise on the instance?
- Do your initial values show up on the right nodes?
- Do enumerations present the names you defined?
Limits
| Lifetime | 2 hours, then it stops |
| Concurrent servers | 1 on the free tier, 3 on paid tiers |
| Pool | Shared and finite — if it is full you are asked to retry shortly |
The lifetime is not negotiable from the UI: start another when you need one. A server is disposable by design, so nothing is lost.
Requirements
Signing in is required — a server is started on your behalf and counted against your account. See Overview.
For a server you control, opcua-modeler generate --server does the same thing
on your own machine, with no time limit — see the
CLI quickstart.