The Vibe Codeability Database

How a Vibe Score is built

• Every app's sheet lists which components its specific build prompt requires — not every feature the real product has, just what our prompt actually builds
• More components isn't automatically worse — a login screen (92) barely dents an average; end-to-end encryption (28) drags one hard

The full mechanism — how ratings combine, who scores, and the honest limitations — is on how verdicts work.

What the score can't see

• Pure network effects genuinely can't be captured here, and that's not a gap to fix — a wiki or a CRM's value depends on other people already using it, which has nothing to do with how hard the software is to build. Notion is the clearest example: buildable components, still a JUST PAY, because good code doesn't manufacture a network
• Everything else started as a gap and got fixed as real apps exposed it — legal weight, licensed content, rendering engines, integration ecosystems, and no-code editors are all real, scored components now. If a future app exposes another one, it gets added the same way, and past apps get retroactively tagged rather than left with an inflated score
• Component interactions still compound beyond what a simple average shows: OAuth plus realtime collaboration plus a marketplace is harder than the three numbers averaged suggest