Three public pages on this site — the Showcase, the Unity Lab, and the Blender Studio — exist for one reason: claims about AI build quality are cheap, and we'd rather show receipts than write adjectives. This chapter explains what each page is and how to read it skeptically, which is the intended way.
What each page shows
| Page | Contents | Read it for |
|---|---|---|
| Showcase | Featured builds across every type, with live previews | Range — sites, apps, games, the variety of art directions |
| Unity Lab | Real Unity 6 builds, each with concept art, engine screenshots, playable WebGL | The ceiling — what the 3D pipeline produces |
| Blender Studio | The 3D asset work by itself: scripts, renders, audit rounds | The process — how assets converge on their concepts |
The part that makes it honest
The lab pages publish the iteration rounds — the early renders where the model looked wrong, next to the round that fixed it. That's deliberate and a little uncomfortable, which is why it works: a curated reel proves someone can curate; visible iterations prove a pipeline converges. When we say builds are audited against their concepts until clean, the labs are that claim wearing its receipts.
These are the loop's own products
Nearly everything on those pages was built by the platform's own autonomous loop — the same research → build → verify → publish machinery user accounts run, pointed at our own product. We eat the cooking; the labs are the plates.