Video Neuro-Evaluator

A scientific gate that scores a video script or short/long video draft from a neuroscience + psychology + sociology + MrBeast retention structure perspective across 12 dimensions (A1–A12) on a 0–10 scale. It scores not on "it felt good," but on the basis of an observable trigger — which sentence, which second, which mechanism.

What does it do? It measures hook, curiosity gap, cognitive load, open loop (Zeigarnik), narrative transportation, emotional arc, pacing, social proof, parasocial bond, identity/belonging, payoff escalation, and TR cultural resonance. It ties each score to a real scientific source (fabricated citations are forbidden) and penalizes manipulative tactics (false urgency, clickbait-and-switch, fear exploitation) with an ethics flag — it does not reward them.

When is it used? Before a video/Reels/Shorts script goes live; as a mandatory quality gate in the content pipeline; in the ai-otorite production line, every script automatically passes through here (≥7.0 → passes).

Output: A plain-Turkish score card — total score + verdict (PASS / REVISE / REWRITE), a justified table across 12 dimensions, concrete rewritten correction sentences for the 3 weakest dimensions + the science behind them, an ethics flag, and a machine-readable JSON block that the pipeline parses.

It applies separate weight sets for long-form (8–15 min) and feed/short (≤60–90 sec); for shorts the hook weight is markedly higher.