Generative AI Plugin for Unreal Engine (C++ Study Project)

What It Does?

The plugin connects Unreal Engine NPCs to:

🌐 Online models like OpenAI’s GPT (via API)
💻 Offline local models using llama.cpp (runs on the player’s machine)

LINK TO MY POST (Also describes my steps how I built it)

The Example Video:

Why This Was Challenging ?

Complex Integration - Connecting Unreal Engine’s C++ systems with external LLMs (both cloud-based and local) required deep knowledge of both Unreal’s architecture and how LLMs work. It wasn’t just a plug-and-play setup - I had to handle performance, memory, threading, and error handling carefully.

Multiple Backends - Supporting both online and offline models added extra layers of difficulty. Llama.cpp works very differently from an online API, so I had to design a flexible system that could switch between them.

Customizable & Usable - I wanted the plugin to be useful to designers (not just programmers), so I focused on making it easy to configure, with clear options and settings inside Unreal Engine.

What I Learned

Go Back

BUAS