Unreal Engine 5.8 gains Unreal MCP plugin for direct LLM integration

Epic Games introduced the Unreal MCP plugin for Unreal Engine 5.8 during its State of Unreal broadcast. The tool reportedly enables large language models to connect straight to the engine, interact with its systems, and carry out tasks based on text instructions.

Epic unveils Unreal MCP plugin during State of Unreal 2026. Senior director of Epic Games' Research & Development department Michael Lentine presented the plugin on 2026-06-17 and confirmed it is already available for Unreal Engine developers. He repeatedly emphasised that creators retain final authority over the output.

Lentine noted that while models handle broad changes effectively, developers excel at specifying precise requirements. The live demonstration divided the display between the running engine and a Claude interface.

Demo shows LLM generating and editing furniture via prompts. Instructions sent to the model produced and modified furniture inside an otherwise vacant space. The session alternated between automated results and direct human adjustments.

Subsequent segments depicted construction of an entire city environment that merged manual modelling with AI assistance. Prompting the system for overcast lighting conditions yielded an inaccurate outcome, which Lentine pointed out on screen.

Manual corrections and hazard implementation demonstrated. The operator fixed the lighting mistake by typing further directions into the chat window. A later demonstration illustrated implementation of a proximity-triggered hazard along a city street.

Lentine characterised the hazard setup as something that would normally involve days of iteration across several specialist teams. The plugin reportedly allowed the team to address it far more rapidly.

Artists complete complex scene in days instead of months. "What we've shown here today would take months to build by hand, but with the MCP server and Unreal, our artists were able to make all of this in days," Lentine said. "When you lower the technical friction, you can iterate more and make better games." The presentation illustrated how generative AI can merge with traditional development methods at a time when many studios already employ such technology as a support tool despite continued industry resistance.