Procedural generation (PCG)
AI & gameplayBuild PCG graphs for procedural generation.
What it’s for
This area builds and edits Unreal’s Procedural Content Generation (PCG) graphs. Your assistant can create a graph, add nodes, and connect them into a working procedural pipeline. It reads the current graph first, then makes changes one node at a time.
Ask for it when you want a PCG graph built or changed without clicking around the graph editor: a points source, a slope or noise filter, a pruning pass, or another graph used as a subgraph. Your assistant places each node at a position in the graph, since there is no automatic layout.
The most common node types, such as slope to density, spatial or attribute noise, subgraphs and self-pruning, have dedicated support, so your assistant gets their settings right. For larger changes it can work with the whole graph as readable text, which you can review and diff.
Limits
Dedicated support covers only the most common node types so far.Things you can ask
Create a graph with a points source
You get a new PCG graph with a node that generates points.
You: Make a new PCG graph called PCG_Scatter under /Game/PCG and add a Create Points node.
Created /Game/PCG/PCG_Scatter with a Create Points node.Connect two nodes
The node’s output feeds the graph output, so the pipeline produces something.
You: In PCG_Scatter, connect the Create Points node’s Out pin into the graph output.
Create Points now feeds the graph output in PCG_Scatter.Add a spatial noise filter
A noise node with the type and seed you asked for joins the graph.
You: Add a Perlin spatial noise node to PCG_Scatter with seed 42.
Added a Perlin2D spatial noise node to PCG_Scatter with seed 42.