Easy to make. Hard to see.
19,000 games launched on Steam last year — one every half hour, and half were never seen. AI only floods the shelf faster. Yet every tool that judges a game looks at it after it ships, once the money's spent. GameLoom moves the answer to the front: build it, fix it, or rethink it — before you commit a year and a half — right inside the tools you build in. First cut in the new mainstream voice (no jargon, hook-first, the infinity loop as hero). Narrated (ElevenLabs · Sarah).
Provenance, not vibes
Start with the answer: a gameloom verdict — from Launch Ready to Reimagine. Underneath sits a scraped warehouse (~370k games / ~1.6M reviews), refined through a provenance-walled pipeline into named comparables and benchmark percentiles — where every score cites the exact evidence behind it, pinned to an immutable snapshot. Size isn't the moat; the discipline is. Narrated (ElevenLabs · Sarah).
The verdict where you build
gameloom's verdict inside your IDE. The MCP answers with comparative honesty — "resembles N games: X graduated, Y died" — and deliberately withholds the prescriptive go-or-kill call until a validation gate proves it beats a baseline. Only the defensible layer ships: the dead games you resemble. Honest about what it won't claim yet. Narrated (ElevenLabs · Sarah).
The hero visual system
The gameloom vocabulary in one shot: the panel star-field, the brain / engine kernel, the green infinity loop that is the closed learning loop, DNA links to named comparables, and the score ring filling to a verdict. Pure-3D, sound off — the system every cut is built on.