7 PROMPTS THAT STOP THE SCROLL

COPYscene: [frame 1 composition — the hook moment, described as a still] hook (0-2s): [the scroll-stopper — what happens in first 2 seconds] body (2-6s): [the payoff — reveal, explanation, escalation] closer (6-10s): [the final beat — loop back, CTA, or visual punch] camera: [ONE move that supports all 3 beats — usually static OR slow push-in, never stacked] lighting: [one setup, high-contrast] color: [punchy, saturated, short-form-optimized] audio: [hook audio 0-2s, body audio 2-6s, closer audio 6-10s — spec each] mood: [attention-grabbing / direct / surprising] duration: 10

COPYscene: [pattern-breaking visual — black frame / upside-down / extreme close-up on unexpected detail] hook (0-2s): the disorientation moment — bass tone, no visual content, single sound cue body (2-6s): the resolution — slow rotation, pull-back, or reveal of [subject] closer (6-10s): [product / brand / CTA] lands clean camera: [the move that resolves the disorientation] lighting: pure dark transitioning to [reveal lighting] color: [matches reveal — high contrast, neon noir, etc] audio: silence or low rumble during hook, full audio on reveal mood: disorienting → satisfying duration: 8

COPY[Shot type] of [character] in [setting], looking directly into camera, [light source]. [Character] says, "[HOOK LINE — under 80 chars, under 2 seconds to say]." [Continuation beat or pause]. [Character] says, "[payoff line]." Fixed camera, [50mm f/2.0 / cinematic style]. [Color grade], subtle film grain. [Mood — direct, grounded, slightly confrontational]. Natural room ambient, no music.

COPYscene: [product on clean surface, top-down or 3/4 angle, branded environment] hook (0-2s): [pattern interrupt OR contradiction — text flashes "wait for it" / object covers product / unexpected motion] body (2-6s): clean reveal of [product] — single motion (lift / pour / unwrap / open) closer (6-10s): macro push-in on key feature, on-screen text "[hook claim]" + "[CTA]" camera: static through hook, push-in for closer, 50mm feel, f/2.8 lighting: warm overhead practical, soft fill color: [brand palette], cinematic, slight grain audio: silence 0-1s, [product sound] at 1s, [whoosh] at 2.5s, [ambient] through body mood: appetite-triggering / desire-creating duration: 10 hard rules: only [product + minimal props] in frame. Do not invent features not in the input image.

COPYscene: [closed box / wrapped product] center frame, premium surface, branded environment hook (0-2s): hands enter frame, single dramatic sound cue (knife slice / tape rip / lid lift) body (2-6s): slow methodical reveal — wrapping pulls back, product emerges, single 360 rotation closer (6-10s): final hero shot of product at rest, on-screen text "[product name]" + "[where to buy]" camera: locked tight on box, follows the reveal motion, ends on macro detail lighting: dramatic single-source key, deep shadows, premium product feel color: dark moody with one accent color from the product audio: silence then sharp open sound, paper rustle, soft ambient through reveal, no music mood: anticipation → satisfaction duration: 10

COPYscene A: [first composition — the setup] scene B: [second composition — the payoff] transition: at 2-second mark, fast horizontal whip-pan from scene A to scene B, motion blur during pan, continuous audio, single take feel camera A: [static or subtle move] camera B: [matches A's energy on resolve] lighting: continuous between scenes, one direction color: matched grade across both scenes audio: continuous through whip, slight whoosh on pan, no cut audible mood: kinetic, premium production value duration: 6

COPYUsing input image 1 for [character name]'s face/body identity. [scene + action prompt as normal] Hard rules: - Character must match input image 1 exactly: [hair color], [skin tone], [eye color], [build], [defining feature] - No [list features that AI commonly drifts toward — e.g. "no blonde hair, no pale skin, no blue eyes, no professional makeup"] - Outfit must match: [exact wardrobe — colors, type, no logos unless specified] - Same lighting direction as input image - Same age range as input image