Geometry and add-ons

Every node in this series so far has been a plain cuboid, cylinder or sphere. That was deliberate, to keep the focus on the model underneath. This last page covers the two things that sit outside the model itself: what a node actually looks like, and how a completed result reaches something other than the built-in plot.

Real geometry instead of a primitive

A node's Appearance field isn't limited to the three built-in primitives. Import file accepts your own STL or STEP geometry, and Shape library offers a bundled, searchable set of ready-made parts.

Selecting a shape shows a live, rotatable preview before committing to it. Applying one sets the node's appearance to "Imported geometry" immediately, visible back in the builder as a plain status line, same as a manually imported file.

Creating the node places the real geometry on the canvas, not a placeholder standing in for it.

This is presentation only. Nothing about a node's shape, imported or built-in, changes how it behaves in a simulation. That's entirely governed by its states and equations, covered everywhere else in this series.

Opening a result somewhere other than the built-in plot

Once a result exists, an add-on toolstrip appears alongside the usual run controls: bundled visualizers with permissioned access to the current result, running in their own window rather than a panel inside the main one.

This is a genuinely separate OS window, with its own titlebar and its own minimize/maximize/close controls, plus a signal picker and comparison plot built for looking at more than one node's data at a time, which the built-in per-node plot ("Exploring results") doesn't attempt to do. Scrubbing its timeline moves the main project window's own time cursor too; they stay synchronized rather than showing two independent views of the same run.

Closing note

That's every page in this series: composing a model by hand, organizing it with subsystems and edge groups, running and reading back a result, describing changes to the model assistant, recovering structure from data instead of authoring it, and finally, what a node looks like and where a result can go beyond the canvas itself. Each page stood alone, but together they cover the whole path from an empty canvas to a result worth trusting.