# VIDEO-PROMPT.md — "Origin" as an AI-generated video sequenceFor text-to-video tools (MiniMax/Hailuo, Kling, Runway, Luma, Veo, or similar). Most of these generate clips in 5–10 second bursts, so the sequence is written as **five separate shots** meant to be generated individually and cut together, rather than one continuous prompt — trying to force a 20-second continuous camera move into a single generation almost always breaks down into warping and drift. Reuse the **Master style block** below, word-for-word, at the start of every shot's prompt — that repetition is what keeps the lighting, palette and material consistent across separately generated clips.## Master style block (paste at the start of every shot)*Cinematic 3D motion graphics render, real-time WebGL / Three.js aesthetic, ultra-clean and minimal, high-end tech brand film. Subject: hundreds of thin, glowing line-segment fragments — like luminous filaments of wire or light-drawn brush strokes — floating in a vast black void. Material: emissive, unlit, faint additive bloom around every line, no visible surface shading, no ambient occlusion, pure light-on-black. Palette: near-total black background (crushed to pure black, no visible haze or dust), fragments in a cold graphite-grey when dim and a bright cobalt-blue-white (electric royal blue with a hot white core) when lit. Mood: precise, engineered, quietly confident — the feeling of watching a hand-built object assemble itself, not a chaotic explosion.*## Shot 1 — Genesis (5s)**Prompt:** In a vast black void, hundreds of short, thin glowing line-fragments drift slowly and independently at different depths, like scattered filaments of light suspended in zero gravity. Most are dim graphite-grey, nearly swallowed by the darkness; a few pulse faintly brighter at random. No shape, no pattern — quiet, sparse, unresolved. Camera holds a wide static shot, slightly high angle, looking into the depth of the void. Subtle depth-of-field: fragments closest to camera are crisp, the deepest ones soften and fade into black (fog falloff). Very slow, barely perceptible drift — no fast motion, no camera movement. Cold, still, anticipatory mood, like the moment before an idea exists.**Camera:** Static wide shot, 35mm-equivalent, eye-level to slightly high, f/2.8 shallow-ish depth for the deepest layer only.**Lighting:** No practical light source — every value comes from the fragments' own emissive glow. Zero fill light; blacks stay pure black.**Motion:** Slow, floating, near-imperceptible drift on independent axes per fragment — think embers in still air, not wind.## Shot 2 — Assembly (6s)**Prompt:** The same scattered glowing line-fragments begin accelerating and converging from all directions in the black void, streaking inward on curved paths like light being drawn by an invisible hand. As they arrive at their destination, they snap into place with a sharp, satisfying settle — no bounce, no overshoot — brightening instantly from dim grey to hot cobalt-blue-white on arrival. The fragments assemble into the bold, geometric wordmark **"MANIAC"**, built entirely out of these short glowing dashes rather than a solid typeface — a constellation that reads as a word only once complete. Camera slowly pushes in toward the forming word as it completes. Final frame: the word fully assembled, glowing crisply against black, faint bloom haloing every stroke.**Camera:** Slow dolly-in, starting wide (as Shot 1 ended) and tightening toward center frame, 35mm-equivalent, locked axis (no rotation).**Lighting:** Entirely emissive/self-lit; brightness of each fragment is the only "exposure" cue, ramping from ~15% to 100% as each piece locks in.**Motion:** Fast converging streaks with hard, snappy arrivals (ease-out, no elastic bounce) — the tempo should feel like typesetting, not fireworks.## Shot 3 — Expansion (5s)**Prompt:** The glowing line-fragments that formed the wordmark begin pulling apart and stretching outward and upward, re-forming from flat 2D strokes into four tall, narrow vertical clusters of light — like glowing structural pillars or light-columns — arranged in a wide ring. Nothing new appears; the same strands of light are visibly repositioning themselves, stretching like glowing threads being redirected mid-air. Camera pulls back and rises slightly as the ring reveals itself below and around the frame, opening up to show the full circular arrangement of four luminous pillars standing in the black void, each humming with a faint inner pulse.**Camera:** Combined dolly-back + slight crane-up, revealing scale; ends on a three-quarter high angle looking down at the ring, 24mm-equivalent for a wider reveal.**Lighting:** Same emissive-only rule; add a very subtle blue rim-glow bouncing off the void floor directly beneath each pillar, barely visible, to sell weight and grounding.**Motion:** Elastic, stretching deformation — lines visibly travel and reshape rather than cut or teleport; medium-slow pace, deliberate.## Shot 4 — Orbit (6s)**Prompt:** Camera slowly and smoothly orbits around a ring of four tall glowing light-pillars standing in a black void, sweeping roughly three-quarters of the way around the circle in one continuous arc. As the camera passes each pillar, that pillar brightens momentarily from cool dim blue to hot white-blue, as if activated by the camera's attention, then settles back to a lower ambient glow as the camera moves on. A faint glowing ring-line traces itself along the ground beneath the pillars as the orbit progresses, connecting each one in sequence. No text, no words — pure light and motion; any labels will be added afterward in post-production.**Camera:** Continuous orbital dolly (a smooth, motion-controlled arc, like a drone or robotic camera rig circling a sculpture), constant radius, slight downward tilt, 35mm-equivalent, no handheld shake, no speed ramping — even, machined smoothness.**Lighting:** Emissive pillars as the only source; each pillar's brightness should visibly rise and fall as the camera's angle to it changes, like a spotlight following the lens.**Motion:** Smooth constant-speed rotation with gentle ease at the very start and end of the arc; the floor ring draws in progressively, synced to orbit progress.## Shot 5 — Convergence (5s)**Prompt:** The four glowing light-pillars begin dissolving and pulling inward, their light-strands streaking toward the center of frame and reforming into the clean glowing outline of a smartphone — a tall rounded-rectangle frame drawn entirely in the same cobalt-blue-white light. As the frame completes, a soft light fills the interior of the phone outline like a screen powering on, and the frame settles, fully formed, glowing steadily at the center of the black void. Camera pushes in slowly to frame the phone outline centered and upright, ending on a clean, symmetrical, resting composition. Leave the interior of the phone as a simple soft glow, not a rendered app screen — the real product screenshot will be composited in afterward.**Camera:** Slow, steady push-in ending dead-center and level, 50mm-equivalent for a tighter, calmer final frame; camera comes to a full stop.**Lighting:** Emissive only, with the phone's interior introducing the first "warm" soft fill glow of the whole sequence — a subtle shift signaling arrival/completion.**Motion:** Converging streaks matching Shot 2's arrival language, but calmer and slower — this is a resting beat, not a climax.## Negative prompt (use on every shot)*No text or legible words appearing in-frame (aside from the intended wordmark in Shot 2), no human figures or hands, no camera shake or handheld jitter, no lens flares, no color outside the blue/grey/black palette, no smoke or visible fog haze, no motion blur trails longer than a few frames, no bounce or overshoot on arrivals, no daylight or warm ambient light, no glitch/VHS/scanline effects.*## Production notesGenerate each shot separately, keep the Master style block identical every time, and if your tool supports a seed or reference/start-frame image, carry the **last frame of one shot as the start-frame reference for the next** — this is the single biggest lever for visual continuity between clips. Don't ask the model to render the pillar labels or the phone's screen contents; AI video models render legible text and UI unreliably, so add "Brief / First look / Your feedback / Approval & launch" as clean text overlays in your editor during Shot 4, and composite your real Discipline Maniac screenshot into the phone outline during Shot 5. Cut on the pillars' settle beat and the phone frame's completion beat rather than mid-motion, and add a single low, rising drone/whoosh under Shots 2–5 with silence in Shot 1 — sound sells the "assembly" moment more than any visual can.