Prediction Error

Pre-Generated Content:

We utilized generative AI tools (such as Midjourney and Stable Diffusion) to assist in creating 2D concept art, UI elements, PBR texture maps (base color/normal/emission), and background skyboxes. All generated assets were manually edited, composited, and integrated into the game engine by the developer.

We used text-to-3D AI models (such as Hunyuan 3D) to generate baseline topology for certain environmental props, which were then manually optimized and textured.

Voice lines for specific in-game entities were synthesized using open-source TTS and voice cloning algorithms (e.g., GPT-SoVITS) based on legally permissible datasets.

Live-Generated Content:

None. The game does not generate any AI content at runtime.