Bring the Soul of Local AI
to Your Unreal Projects
Transcending traditional source code comments, we delve deep into physical logic. LiteRT-LM Unreal is a high-performance local Agent inference solution designed for AAA-grade projects.
ABI Firewall: Zero-Deployment Integration
We establish a physically isolated ABI firewall through litert_lm_wrapper.dll. This means that when integrating 70+ low-level dependencies (Bazel, Rust, TensorFlow, WebGPU), you don't need to install any complex build environments on the host machine. The plugin is plug-and-play.
Tech Stack Glimpse
Physical KV Hits
Based on low-level C-API Session pointer management, achieving instant Agent switching with KV Cache hit rates exceeding 98%.
Extreme VRAM Control
Fully automated VRAM budget allocation. Supports hot-cold VRAM swapping, running 7B-level multi-Agent conversations smoothly even on 8GB VRAM.
Download Demo Project
Want to see it in action immediately? Get the full open-source demo from GitHub, including all interaction logic and example assets.
Buy from Fab
Current phase: FREE (Price will increase to $19+ approximately before 2026.5.30). Get it now for lifetime free updates.