The key idea: every node in a scene graph can be expanded into its own detailed sub-graph, infinitely. Then Gemini's image model (Nano Banana) renders each level as actual artwork (you pick your style).
Requires a GCP account for image rendering — I know it's friction, but Vertex AI was the only way to get reliable image generation.
Feel free to use for your own projects or implement the concept better than I did.