The plugin runs within Poser. It has its own viewport where the render occurs. You can set the viewport to react to changes in the poser scene (sun, camera changes, prop changes, morphs, poses, etc). But this render viewport is just what the poser camera sees.

The VR portion is a panoramic camera which is not present in Poser. Even if it were, it would not properly display on a screen. So you can set up your scene in Poser, get live updates in the Octane viewport, but when you want to watch in VR, you need to render it to a panoramic camera. The viewport changes to a 2d view of that ,where the up/down/le/ri for each eye are rendered next to each other , which is what the occulus format essentially is.

To see the VR output you need to view it through a VR display after the render is finished.

No other 3rd party apps are involved . The only changes you have to make to your poser scene are the materials (if you are not satisfied with the default conversion). If you make any changes, they are preserved in the Poser scene, so next time you load your scene, they are all there and you can continue where you left off.

To be honest though, I have not made any animations in VR, so whether that works properly, I do not know.