The Unreal MCP server provides LLM-driven access to Unreal Engine for Fortnite (UEFN) and can interact with a variety of engine systems and editor workflows.
This guide focuses on the dedicated Procedural Content Generation (PCG) toolset you can access through the MCP server. These tools are designed to manipulatePCG systems and graphs directly using LLM interaction. PCG workflows involve complex graph structures, procedural logic, and attribute-driven data flow, it is important to follow specific practices.
This document recommends workflows and practical usage patterns gathered from day-to-day production use.
Prerequisites
Enable the Unreal MCP and start the MCP server. For more information on starting the server and connecting your LLM, see UEFN MCP.
With the server running, ask your LLM to read the PCG graph generation skill. Your LLM discovers the toolset on its own; the skill gives it the foundational understanding it needs to work with PCG graphs reliably.
Reading the skill is a required part of session setup. Without it, the LLM will often:
Misunderstand PCG concepts
Overcomplicate solutions
Misuse nodes or parameters
Produce unreliable graph logic
Skills are assets in your project, and your LLM reads them through the MCP server. Ask it directly, for example: “List the available skills, then read Skill_PCGGraphGeneration.”
Skill_PCGGraphGeneration
This is the primary skill for tasks that involve generating content through repeatable rules or spatial computation. For example, scattering vegetation, populating city blocks, or building road networks. The skill has access to the full library of primitive subgraphs, the most common native nodes, and the example graphs, and how to wire them together correctly. For less common nodes, the LLM can look up a node's pins and parameters as it works.
Use this skill when the task can be expressed as a system or a pattern, not a one-off placement.
Key Principles
The LLM performs best when:
It is properly grounded in context.
It has analyzed existing examples.
It is given explicit context.
It is kept within well-defined tasks.
It is iterated through tight feedback loops.
It is encouraged to reuse and adapt existing solutions whenever possible.
Recommended Workflow
The most effective LLM workflows in UEFN are:
Collaborative between you and your LLM
Reference-driven
Iterative
When using the Unreal MCP server with the PCG toolset, use the following workflow:
Check that the Unreal MCP server is running.
Ask your LLM to read the PCG graph generation skill.
Identify relevant example graphs.
Select relevant assets or actors manually in the Content Browser.
Ask your LLM to review references before planning.
Ask for a plan or graph review before execution.
Execute incrementally to provide oversight.
Supervise and course-correct actively during execution.
Set Up Your Context
Use Existing PCG Example Graphs
Reference-driven workflows consistently produce the most reliable results.
One of the most effective ways to improve LLM output quality in PCG workflows is to explicitly reference existing example graphs.
This includes:
Example graphs provided in UEFN
Personal project example graphs
Existing production-ready graphs
Reference implementations created by experienced users
Treat these graphs as reference material that the LLM should follow.
Without references, the LLM may:
Overcomplicate solutions
Misuse nodes or parameters
Build unreliable graph logic
Ignore established workflow conventions
For complex tasks, tell the LLM to take the following steps:
Inspect existing examples first
Reuse established patterns
Follow existing graph conventions
Compare proposed solutions against working implementations
Your LLM's example graphs come from the PCG folder, located in Epic > PCG > SampleContent and Epic > PCG > Primitives >Examples. They are organized by themes such as Basics, Buildings, City, Forest, and Road.
Even though this guidance already exists inside the PCG graph generation skill, LLMs may still overlook references and attempt to build solutions from first principles. Reinforcing the importance of references during the session significantly improves consistency and accuracy.
Adapt Existing Patterns
In many cases, it is more effective to reuse or duplicate an existing graph than to build a new one from scratch.
A good general workflow is:
Reuse an existing graph if it already solves the problem.
Duplicate and modify a graph if it is close to the desired result.
Only create a new graph when no suitable reference exists.
This approach preserves proven methods and graph structures while reducing ambiguity. This avoids unnecessary node creation and can significantly reduce context window usage and token costs.
Prompt Examples
“Look for an example graph that performs a similar operation.”
“Use the existing PCG examples as reference before modifying the graph.”
“Find how this node is used in the examples.”
“Match the conventions used in the existing project graphs.”
“Review relevant examples before proposing a solution.”
Use Explicit Context
Direct user-provided context is faster and more reliable than autonomous discovery. When interacting with assets or actors, give the LLM explicit context instead of asking it to discover content on its own.
It is recommended to use the following steps:
Select assets directly in the Content Browser.
Select actors directly in the Level Editor.
Issue the request or action.
This approach is typically faster, more reliable, and uses less context and execution time.
Without explicit context, the LLM has to discover content by calling many read tools in sequence, which slows down the interaction, increases token usage, and adds unnecessary steps. With a selection in place, it reads your selected assets and actors directly through the Editor Toolset.
Plan Execution Steps
Planning and reviewing the execution steps can improve reliability for complex workflows. For complex operations, ask the LLM to explain or review its approach before execution.
This method is useful for:
Large graph edits
Multi-step workflows
Attribute manipulation
Experimental setups
For these requests, explicitly ask the LLM what references, examples, or existing graphs it intends to inspect before it acts. Encourage it to base its plan on existing examples whenever possible.
Prompt Example
“Review the graph before acting.”
“Explain the intended approach first.”
“Provide a step-by-step plan.”
“List the nodes you intend to modify.”
“Base the solution on an existing example graph if possible.”
“What examples or references are you planning to use?”
Monitor Execution
Treat LLM-driven workflows as supervised collaborations between you and your LLM, not as fully autonomous execution.
During execution, it is recommended to do the following:
Monitor actions as they happen
Interrupt incorrect approaches early
Reorient the LLM with references or examples
Provide concrete corrections
If a step produces errors or unexpected results, review the issue with the LLM before proceeding with additional automated modifications.
Shorter feedback loops are more efficient than allowing the LLM to fully complete an incorrect approach.
Expensive Operations
Some PCG operations can become expensive through the MCP interface, in terms of execution time and context usage. For example:
Large graph analysis
Multiple graph analysis
Large dataset inspection
Broad asset analysis
Most PCG tools return their full result, so keep requests narrow: ask about one graph at a time, and when inspecting point data, ask for a specific attribute and a range of points.
Data View
Inspecting PCG attributes means analyzing generated PCG point data and associated attributes, similar to reading a large procedural spreadsheet.
The LLM cannot inspect PCG attributes directly on its own. For attribute inspection, it uses the Data View tool, which returns the output data of one node from the last time the graph ran. The first request only turns inspection on, so the LLM needs to run the graph again before the data is available. To keep the response small, ask for a specific attribute or a range of points. Nodes that run more than once, such as nodes inside loops, aren't supported.
Even though Data View is part of the PCG toolset, it can still be beneficial to explicitly mention it when prompting, as the LLM may not always select it automatically.
Prompt Examples
“Use Data View for the attribute inspection.”
“Inspect the PCG points through Data View.”
“Use Data View to analyze the generated attributes.”
Additional Skills
The PCG toolset includes more skills that extend the PCG workflow. Your LLM can list them on its own, but it helps to ask it to read a specific one when you know it applies.
Skill_PCGShapeGrammarDefinition
Use this skill when the output requires a composed, rule-driven sequence of modules that fit together like puzzle pieces along a path. This includes content such as building facades, corridor modules, street furniture arrangements, or any linear sequence that follows grammar rules.
Skill_PCGMeshPartition
Use this skill when your project uses Mesh Terrain as its landscape system and you need to interact with it using PCG. It explains to the LLM how the terrain modification layers are ordered, how material weight channels work, and how to author PCG graphs that correctly write to the terrain without breaking the priority system.
Skill_PCGInstancingOnMeshActor
Use this skill when you want to cover a mesh with detail, such as sci-fi greebles, surface clutter, and micro-props scattered across an object's surface. It guides the LLM through the Scatter_points_onto_Mesh_Actor graph template, which uses Poisson sampling, and warns it about generating millions of points on a large or heavily scaled mesh.
Skill_PCGBiomeCore
Use this skill when you are creating a biome system from scratch and want to define what assets spawn where based on landscape rules, priority zones, and density filters. It walks the LLM through the full asset creation chain (definition, assets, generator, filter) and placement in the world, with guardrails for the most common silent failures that cause nothing to spawn. The PCG Biome Core plugin must be enabled.
Skill_PCGBiomeCoreRefinement
Use this skill when a biome you already set up isn't behaving: nothing spawns, you get zero points, or only part of the area generates. It gives the LLM a step-by-step checklist for finding the cause.
Skill_PCGAssetZoo
Use this skill when you want to line up a set of assets for review, such as a grid of meshes with frames. It uses the Asset_grid_with_Frame graph template.
Skill_InstantLevelOperations
Use this skill for quick, one-off batch operations in the level that don't need a graph of their own, such as scattering or aligning a selection. The LLM runs an instant graph and moves on.