A set of local tools that turn a prompt into usable game and app assets on our own GPU, instead of paying per generation and waiting on a queue we don't control.
The pieces
A live image-to-3D WebUI drives SDXL reference generation through ComfyUI and streams Hunyuan3D-2 model previews into the browser as the diffusion runs. Alongside it sit a texture-generation batch tool, a game SFX/ambience generator, and an app-icon packer that outputs Android, iOS, and web icon sets from one source image or a generated SVG.
Where it fits
It's internal tooling first: it feeds Aethelgard's low-poly art, PartyFuse's icons, and demos for clients who want "AI-focused but practical." The hard part isn't generation anymore; it's curation, so the next step is tagging the genuinely usable outputs rather than generating more.