Music Video

Realistic live-action dance short video. The character uses the face, hairstyle, body type, and identity features of preset Character #1 throughout. Character #1 is an adult. Wears only the preset outfit throughout, maintaining delicate facial features, wearing decorative glasses, and deep purple high heels. Identity, hairstyle, body, outfit, and all accessories remain consistent. Total length 11.20 seconds, 9:16 vertical screen, 30fps, single continuous shot. Character centered, facing the camera directly, movements light, playful, natural smile. Strictly execute the following dance sequence without adding new moves or extending preparation/hold times. [Audio Lock] Use the uploaded file "lumiraexo_on_Reels_audio only.mp4" as the sole voiceover and background track, starting synchronously at 00:00. Retain original vocals, music, rhythm, tone, speed, and pauses; no re-voicing, pitch shifting, or added music/dialogue. Track ends at ~11.17s, leaving ~0.03s natural silence. Body turns, palm pushes, knee lifts, and hand gestures follow the actual beat of this audio; do not set a conflicting BPM. Maintain natural smile and breathing; do not add obvious lip-sync speech. [Segment-by-Segment Action, Expression, Gaze] 00:00–00:01.10 | Alternating Palm Push: Legs naturally apart, knees relaxed, hips shifting weight left/right on beat. Left hand behind left ear, left elbow open outward; right arm extends toward camera, fingers spread, palm forward, forming a clear "stop" gesture. Then right hand returns to right ear side, left palm pushes forward, then retracts. Smooth hand exchange, palms do not touch lens. Expression: Bright natural smile from start. Gaze: Always direct eye contact with camera. Landing: Each palm push reaches max forward on beat, hips shift simultaneously. 00:01.10–00:02.20 | Hands Behind Ears Rhythm: Following palm push, both hands move behind ears and back-of-head sides, elbows open naturally. Torso relaxed, shoulders slightly rebounding, hips continuously swinging left/right. Head tilts gently then straightens, hair sways naturally with shoulder movement. Expression: Smile softens to playful, focused slight smile. Gaze: Looks at camera first, briefly looks down near end of segment. 00:02.20–00:03.40 | Single Knee Lift: Right hand stays behind right ear, left hand descends along outer body side, stops in front of left thigh. Weight stable on right leg, left knee lifts forward/upward, opens slightly left, left hand hovers above lifted thigh; then left leg returns to ground along same path. Complete one continuous knee lift only, no kicking, no pulling clothes. Torso tilts slightly for balance. Expression: Light lip press, playful seriousness. Gaze: Looks at left leg during lift, back to camera upon landing. Landing: Highest knee point hits beat, landing immediately transitions to next gesture. 00:03.40–00:04.40 | Chest Fingertip Frame: Left foot planted, both hands return to chest, palms angled toward camera; thumb and index finger approach, briefly forming small triangle frame. Then hands expand from center to sides, descend in arc to waist. Body maintains slight left/right rhythm. Expression: From restrained smile to bright smile. Gaze: Looks at central fingertips, raises eyes to direct camera view as hands open. Landing: Fingertip closure and hand opening correspond to consecutive beats. 00:04.40–00:07.20 | Low Hands, Continuous Hip Swing: Elbows slightly bent, hands relaxed into small fists, positioned at body sides, slightly below waist, knuckles facing front-inside. Keep hands low, rely on slight knee flexion/extension for continuous left/right weight shifts, hips thrust shortly through center then switch; shoulders counterbalance slightly, forearms bounce naturally with rhythm. Feet stay mostly in place. Expression: Relaxed, bright smile, cheeks naturally raised. Gaze: Mostly direct camera, occasionally offset with head tilt, immediately returning. Landing: Each hip thrust clearly hits beat, no slow motion, no dragging. 00:07.20–00:09.20 | Maintain Rhythm, Brief Smile Reduction: Continue low hands and left/right hip swing, amplitude consistent with previous segment. Head tilts gently to one side then straightens; shoulders produce fine rebounds with beat, body does not move significantly forward/backward. Expression: Smile naturally recedes, lips close lightly, forming brief calm state; then corners of mouth rise again. Gaze: Briefly looks aside, returns to camera, maintaining interaction. 00:09.20–00:11.20 | Smile Return, Natural End: Continue original left/right weight shifts, hands remain at body sides, wrists relaxed. Head tilts lightly with beat, finally returning to front, hip movements naturally reduce. Expression: Reveals bright smile again, ending in relaxed, happy state. Gaze: Direct camera. Landing: Follows final rhythm of track to settle naturally, no sudden stiff freeze. [Camera & Scene] Fixed phone front position, lens approx. chest height, slightly upward. Vertical medium-full shot, capturing from above head to mid-lower thighs, preserving space for palm push and knee lift. Natural perspective enlargement when reaching forward; camera stays static, no zoom, pan, rotate, or cut. Scene is simple modern interior: Two dark wooden doors behind character with silver handles and closers; surrounding light walls and white ceiling, small circular fixtures visible overhead. Retain everyday feel of real residence or hotel room. Soft indoor ambient lighting illuminates face and body, stable background exposure. Image has authentic smartphone shooting texture, retaining skin texture, hair strands, and fabric inertia, no extra light effects. Exclude: Face swap, identity drift, hairstyle change, outfit variation, disappearance of stockings/glasses, heel changes; multiple hands/fingers, fused fingers, palm-through-face, joint hyperextension, knee-lift imbalance, foot sliding, chaotic leg kicks; action teleportation, mechanical hip swings, rhythm misalignment, fixed unchanging smiles, gaze drift, abnormal lip-sync; audio speed changes, track misalignment, extra voiceovers/music; push/pull camera moves, digital zoom, cuts, background deformation, severe blur, subtitles, watermarks, logos.

Make the lights softly twinkle and make it slow slowly start to snow and make some beautiful piano music

Ultra-cinematic Korean music video dance sequence featuring a beautiful young Korean girl performing stylish K-pop choreography alone on an atmospheric neon-lit Seoul night street. The video should feel like a real high-budget K-pop MV mixed with fashion-commercial cinematography, with continuous dynamic camera movement throughout the entire sequence instead of a fixed angle. The girl wears fashionable modern Korean streetwear: fitted blue jeans, stylish cropped top, layered silver accessories, trendy sneakers, soft natural makeup, and long flowing dark hair reacting naturally to movement, wind, and motion. She has elegant idol-like visuals, charismatic stage presence, confident expressions, and smooth professional dance performance energy. Environment: aesthetic empty Seoul alleyway at night with glowing Korean neon signs, convenience store lights, wet reflective pavement, cinematic haze, soft fog, blue-magenta neon glow, warm street lamps, distant headlights, urban reflections, atmospheric bokeh, and premium K-drama nightlife visuals. The street should feel cinematic, immersive, and stylish without distracting from the performer. The choreography flows naturally like a professional K-pop performance: intro pose, hand wave, hip sway, side-step movement, hair flip, arm extension, spin transition, mid-dance jump, sharp synchronized choreography pose, body roll, detailed footwork, dynamic lean, turning movement, confident stance, ending flourish, final cinematic pose facing camera. IMPORTANT: The camera itself constantly moves dynamically like a real music video production. Use cinematic camera movements throughout: slow dolly-in during intro, smooth side panning shots, tracking shots following footwork, low-angle moving shots, subtle handheld realism, dramatic push-in zooms, cinematic zoom-outs, orbit camera movement around spins, tilt-up shots during jumps, parallax motion between foreground and subject, dynamic reframing, Dutch-angle transitions, close-up to wide-shot transitions, sweeping gimbal movement, lens breathing effects, rack focus transitions, motion-blur camera sways, cinematic follow-cam energy, smooth Steadicam movement, high-fashion editorial framing, music-video-style camera choreography. The camera angles should constantly evolve: wide establishing shots, medium tracking shots, low-angle hero shots, dynamic side profiles, close-up emotional moments, full-body dance framing, cinematic over-the-shoulder movement, dramatic front-facing hero compositions. Visual style: ultra realistic, photorealistic Korean idol aesthetic, premium cinematic lighting, glossy reflections, dramatic rim lighting, volumetric fog, realistic fabric physics, natural skin texture, cinematic motion blur, realistic hair movement, shallow depth of field, highly detailed urban night atmosphere, luxury K-pop commercial look. The entire video should feel like a professionally dire

She performs an energetic K-pop girl-group choreography, perfectly on beat. Sharp, precise moves: crisp arm isolations, quick head flicks, confident hair whip, a hip sway into a step-and-turn, then a powerful point choreography with both arms. Smooth...

Create a beat-synced dynamic graphic K-pop music video for this song. Use this mood board, the K-pop lead singer character design, and the backup dancer and Agent characters. Add motion and lip-syncing for the singers and Agents to bring the entire music video to life. Go all out. <Musical Style> <Lyrics>

dance as if listening to this song "“SHE” by Alice Phoebe Lou live — good choice for the beach drive. Beth’s playlist is solid....

Create a music video titled 'Claude Ban Song'. Handle lyrics writing, composition, vocal synthesis, video production, and editing entirely through code.

Cinematic synthwave with a sleek futuristic edge, driving mid-tempo pulse, deep analog bass, shimmering arpeggiated synths, punchy programmed drums, and wide atmospheric pads. Glossy neon production with dynamic builds, crisp transients, and a cool, polished sense of motion and nocturnal energy. No Singing.

Create an ultra-realistic video of an adult K-pop couple dancing in a professional Korean dance studio. Characters: A beautiful woman with long, wavy ginger-copper hair and two small braids, wearing a cropped jacket, fitted top, cargo pants, and sneakers. A handsome man wearing modern K-pop streetwear, a black jacket, cargo pants, and sporty sneakers. 0–3 seconds: Both perform synchronized shoulder bounces, quick side steps, and sharp hand movements while smiling. 3–6 seconds: They perform energetic footwork, body grooves, coordinated arm movements, and a synchronized turn. The woman's hair moves naturally. 6–8 seconds: They move closer, mirror each other's dance moves, exchange romantic glances, and smile playfully. 8–10 seconds: They perform a synchronized final move and finish in a stylish couple pose, smiling at each other. Music: Upbeat K-pop instrumental with strong electronic beats, energetic bass, and catchy synths at 125–135 BPM. Music plays continuously from 0–10 seconds. No dialogue. Camera: Full-body shots, smooth tracking, cinematic lighting, realistic mirrors and reflections. Quality: Photorealistic 4K, natural movements, consistent faces and outfits, realistic hair physics. Negative Prompt: No cartoon, CGI faces, distorted anatomy, extra limbs, robotic dancing, face morphing, outfit changes, flickering, blur, camera shake, or sudden transitions.

Vertical 9:16 cinematic video, 10 seconds, starring Sydney Sweeney as the main character, with her recognizable facial features, wavy blonde hair, and expressive smile. Ultra-fast pacing, whip-pan transitions, and beat-synced cuts approximately every second. 0–2 seconds: Sydney Sweeney, wearing a sleek black freediving wetsuit with a small action camera attached to her chest strap, swims rapidly through crystal-clear turquoise water above white sand. Sunbeams ripple across her face as the camera tracks her with dynamic underwater movement. 2–4 seconds: She speeds past vibrant coral reefs toward a shimmering curtain of iridescent beads. A dramatic speed ramp follows as she bursts through the sparkling curtain. The camera whip-pans into a mysterious dark coral cave, where violet light illuminates the surrounding water. 4–6 seconds: Explosive reveal of a futuristic underwater nightclub filled with neon lasers, colorful strobes, floating luminous star-shaped sunglasses, and tropical fish moving rhythmically to the music. Sydney emerges through a glowing violet portal as cyan, magenta, and purple lights illuminate her face. 6–10 seconds: Sydney puts on oversized neon star-shaped sunglasses, spins playfully in zero gravity underwater, and laughs as glowing bubbles and particles swirl around her. The camera circles her rapidly with energetic zooms, sharp whip-pans, and smooth speed-ramp transitions, ending on a dramatic close-up of her smiling face against the vibrant nightclub. Visual style: Hyper-realistic cinematic footage, lifelike facial details, consistent character appearance, realistic underwater physics, volumetric lighting, vivid turquoise water, neon magenta and cyan highlights, violet glow, floating particles, smooth motion blur, high contrast, premium music-video aesthetics, 24 fps. Audio: Energetic electronic dance music, deep underwater whooshes, sparkling sound effects, and a bass drop during the nightclub reveal. Continuity: Keep Sydney Sweeney recognizable and visually consistent throughout all shots. No face distortion, duplicate characters, flickering facial features, unwanted text, logos, or watermarks.

Photorealistic cinematic narrative K-pop micro music video, vertical 9:16, exactly the same fictional adult identity/identities and clothing as start image. Natural human movement and stable anatomy, real photographed environment, no text, no logos, no new people, no singing or lip-sync. MAGNETAR. Based on lyrics 'My face reflected in the subway window / a little bolder than usual.' Begin with same Lea in cherry-red leather jacket on the night subway. 0-4s medium shot with physically accurate single window reflection: tunnel lights move, she studies her face, slow breath and shoulders straighten with resolve. 4-7s she looks down at her phone, makes one decisive tap as if choosing to reach out first; expression turns quietly brave. 7-10s camera arcs subtly toward frontal three-quarter close-up, she lifts her gaze toward train doors and gives a confident small smile. Keep reflection synchronized, preserve rose hair and exact face, no fast blur or duplicate person. Intimate city-night romance with cool steel/red lighting. Settings: Seedance 2.0 Fast / 720p / 9:16 / 10-second source / generated audio off. Original Suno track added in FFmpeg.

15s Halloween Zombie Duo Dance | 9:16 Vertical | Two Dancers | Fixed Background | Fixed Camera | Beat-Synced Zoom & Close-Up | Horror Trap Hip-Hop MUSIC / ENERGY: An intense beat that kicks in suddenly right from the start, 155 BPM horror trap hip-hop. Hard drums, heavy bass, High speed, sharp accents, energetic and dynamic. ABSOLUTE LOCK: Exactly two dancers only. Dancer A always stays on the LEFT, Dancer B always stays on the RIGHT. Same faces, bodies, hairstyles, outfits, positions, background, lighting, colors, and shadows throughout. No swapping, no crossing places, no morphing, no extra dancers, no cuts. Background never changes. Zombie Dance (Top Priority): Both dancers extend their arms straight out in front at shoulder height, bend their wrists precisely 90 degrees downward, and spread their fingers naturally and straight, unleashing an incredibly fast, aggressive, and high-energy "zombie dance." The performance features a sequence of movements including rapid "zombie steps" with stiff legs, explosive shoulder actions, intense chest pops, sharp head movements, powerful knee lifts, wide-kneed squats, sudden lunges, and vigorous, rhythmic body shaking. Use DOUBLE-TIME footwork at 155 BPM, rapid acceleration, abrupt freezes, explosive restarts, and hard beat accents. Include some sections where both dancers perform different dance moves at the same time while keeping the permanent zombie arm pose. DUO DANCE RULE: This is a duo dance, but they must not be perfectly synchronized all the time. They move in the same overall rhythm, but with slight timing offsets, alternating accents, and different body actions. FINISH: The two of them strike a zombie pose, and the shot zooms in to a close-up of their eyes glowing red as they stare into the camera. APPEARANCE LOCK — ABSOLUTE PRIORITY: Preserve both dancers' original faces, natural skin tones, hairstyles, and makeup exactly as shown in the initial frame. No zombie makeup, horror makeup, corpse makeup, pale or gray skin, darkened eye sockets, facial discoloration, or grotesque facial effects. No blood, fake blood, bleeding, wounds, scars, bruises, or visible injuries. Express the zombie character exclusively through stiff body movements, unnatural head tilts, dragging steps, sharp shoulder hits, and eerie facial expressions. Do not alter either dancer's appearance throughout the video. CAMERA: Fixed frontal camera only. No camera movement, no pan, no tilt, no dolly. Use only fast beat-synced optical zoom-ins, snap zoom-outs, and brief punch-in close-up framing to create speed and impact while keeping the camera position fixed. Keep the same background at all times.

[Segment D: Original 26.667-35.900s; Generate 10s, keep first 9.233s for editing] Seedance 2.5, reference-to-video, 9:16 vertical screen, generate 10 seconds, enable audio output. 3D film-grade CG stadium large-scale concert, generate all shots of this segment at once: lead singer, drummer, bass drum pedal, bassist, lead singer. Normal speed, cut by time, do not compress dwell. Audio runs continuously through shots. Reference binding: @Image1: Dragon Girl, lead singer and electric guitarist, silver-white high ponytail, cyan forehead ornament, cyan-green-white outfit, red laces and gold accessories. @Image2: Shinobi, bassist, golden high ponytail, headband, purple outfit, green cloak and gold shoulder ornaments. @Image3: Female Police, drummer, purple short hair, white hat, glasses, blue uniform, black shorts and cloak. Each image is a multi-view character sheet. Maintain original face shape, body proportions, clothing, and color scheme. Do not generate collages or duplicate characters. Do not copy main characters into the audience. @Audio1: Dedicated 10-second song audio for this segment, starting from original 26.667s; as continuous main music and existing vocals, not just rhythm/timbre reference. Venue, Positioning, and Lighting: Nighttime open-air stadium, giant main stage spanning one end, tens of meters wide, thick platform base; oversized central LED main screen several times taller than figures, wing screens, tall lighthouses, and top steel frame light arrays extending to screen edges. Screens play blue-purple abstract music waves, no text, logos, or figure images. Field full of audience, both sides multi-layer curved stands covered with continuous light sea. Normal positioning: Dragon Girl at front center mic stand, Female Police at back left elevated drum kit, Shinobi at back right. Dragon Girl uses white-cyan-green-gold 6-string electric guitar, Shinobi uses deep purple 4-string bass, Female Police uses royal blue drum kit, silver stands, and gold cymbals. White spotlight illuminates faces and hands, rows of blue-purple beams shoot into night sky, warm gold rim light outlines hair and hem; thin fog concentrated on stage, not blocking stands. Natural proportions between figures and instruments, surrounding stage structures significantly larger than figures. Time starts from 0: 0.000-2.233s, Lead Singer: Dragon Girl frontal slightly low angle medium close-up, face and mic stand on central axis, upper part of guitar at bottom edge. Right hand lightly holds below mic on stand, left hand supports neck, strap stably supports guitar, no strumming yet. She follows audio track to unfold singing phrase, then naturally looks up, eyes engaged. Camera has only restrained slow push, white spotlight illuminates face, partial giant screen and light beams in background. 2.233-3.867s, Drummer: Direct cut to Female Police frontal upper body close-up, drum rim occupies bottom, hands and sticks visible. Alternating strikes, sticks rebound after hitting, cymbal resonance, body small nodding beat, focused expression clear. Elevated drum kit and giant screen positions continue same venue. 3.867-5.267s, Bass Drum Pedal: Single hard cut to low angle side close-up inside drum kit, only showing Female Police's one foot wearing original black high heels, pedal, linkage, beater, and drumhead. Front sole stably lands on pedal, ankle presses down with beat; pedal descent drives beater to hit drumhead, then rebounds. Use this continuous mechanical process to express drumming power, stable shot, no second duplicate foot, no male boots swap, no brand text. 5.267-6.800s, Bass: Direct cut to Shinobi slightly low angle half-body close-up, face, bass body, neck, and hands clear. Body naturally leans forward, left hand changes position on fingerboard, right hand continuously plucks, looks down at neck then gently nods. Camera moves slightly closer, keeping complete hand action, background is tall stage light racks and blue-purple giant screen. 6.800-9.233s, Lead Singer resumes playing: Direct cut to Dragon Girl frontal slightly angled waist-up medium close-up, right hand resumes strumming, left hand fretting. She first looks down at neck, then raises eyes to mic continuing synchronized singing. Camera slowly approaches, retaining guitar body, hands, and face, mic does not cover entire face; rear musicians slightly blurred but correctly positioned, lighting coherent. 9.233-10.000s: Only continue Dragon Girl playing/singing as editing buffer, do not pull out to wide shot prematurely. Sound: (@Audio1 as sole main music and existing vocal source, plays continuously from 0s at original speed, all cuts maintain song time progression; audience shots also continue same music, no restart, no new BGM, no intro added. Lip sync, guitar strumming, bass plucking, and stick landing follow reference track. Do not synthesize additional singing, dialogue, narration, or start/end sound effects. Ambient sound slight, lower than main track, not covering song.) Consistency and Exclusions: Fine 3D film-grade CG materials, normal skin tone, vivid expressions, retain dark parts of original clothing, no dark makeup. Fingers contact strings and fingerboard, sticks rebound after hitting, hair/hem move naturally only. Do not generate small venues, bars, small round stages, empty stands, sparse audiences. No extra limbs, duplicate characters, instrument deformation, clipping, identity swap, clothing drift, skeletons, spiderwebs, ghost faces, magic circles, or combat spells. No subtitles, text, logos, watermarks, character tags, or game UI.

A cinematic, dreamlike fantasy music video featuring two graceful young Asian women in ethereal, celestial outfits. One woman wears a fluffy white textured winter hat, pearl earrings, and a white fluffy dress while holding a fluffy white sheep and a white rabbit. The other wears traditional-inspired blue embroidered nomadic attire with long flowing ribbons. Scenes transition through a surreal, magical landscape: walking on a glowing white stairway descending through clouds and white daisy fields, riding a vintage bicycle together through a vast field of white daisies under a twilight blue sky, dancing barefoot surrounded by billowing white translucent fabric, sitting by a serene dark blue lake shore under moonlight, and gently placing a white daisy in each other's hair. High-end commercial production value, soft volumetric lighting, dreamy atmospheric fog, magical glitter particles floating in the air, emotional and whimsical mood, shot on 35mm lens, 4k resolution, cinematic color grading.

Make a fully interactive music video for this song as a single HTML file, at the level of a motion designer, sound engineer and creative director's showreel. The first frame has to stop the scroll and hook the viewer. Go all out. Lyrics are below. Sync them word by word to the actual vocal. Separate the stems first, then time every word from the vocal track rather than estimating. Detect tempo, beats, kicks and snares, and drive the visuals from them. Rules: - One visual and type system across every section. Quiet sections should feel like the same video with the energy held back, not a different style. - Every word must be crisp and readable over the effects. Glow and trails go behind the type, never through it. - Give each lyric its own visual idea where it fits. No section should be filler. - Don't add any audio to the song unless I ask. - Interactions: hold to muffle the band, tap for a spark quantized to the beat, scroll for gain, stem mute buttons, a scrubbable timeline. - Include an "Export 1080p video" button that renders every frame on a fixed clock and saves a 1920x1080 16:9 MP4 (H.264 + AAC) with the audio muxed in. - Make the HTML file run on its own by double-clicking it in Chrome, with the audio built in. [LYRICS WITH SECTION TAGS]

Bring me to life by Evanescence and have Amy Lee singing on a balcony over looking fire and her band in background make it dramatic, lets say dark smoky elegant and gothic balcony

Create a 50-second vertical music video using programmatic rendering (Three.js) instead of native video AI. 1. Generate song with Suno. Extract lyrics from embedded metadata. 2. Separate stems with Demucs (vocals, drums, bass, other). 3. Transcribe vocals with mlx_whisper. Align with official lyrics to create script where '1 lyric line = 1 scene'. 4. Write storyboard: define 'What to show in 3D', 'Camera', 'Main word' for each lyric line. 5. Have 4 parallel Opus sub-agents write scene programs. 6. Export 16-frame contact sheet for visual verification. 7. Render headless Chrome at 1080x1920, 60fps. 8. Normalize volume to -14 LUFS. Style: Engraving-like lines, colors limited to Black, Bone, Orange.

<inputs> Ask me for: a voiceover or song with a spoken line (the audio file) and its transcript, 10-12 small objects that sum up my life, two silhouettes to light like a heat camera (a head in profile and a raised hand), and the 3 words I want to land hardest. If I skip any, use the defaults: vinyl, book, camera, cat, coin, game controller, cap, heart, money, plant, clapperboard, skateboard; free CC0 silhouettes from Openverse; action, intention, curiosity. </inputs> <direction> A 20 second 4:3 film, 1440x1080 at 60fps, cut to the voice. Off-white paper, black, one red, one yellow, Geist for every word. Film grain on everything, a soft vignette, and a camera that never rests: a slow handheld float under everything plus a slow push inside every scene. No scene cuts to the next: each handoff starts on the exact last frame of the scene before. The only hard cuts are the beat flips on the last word. Every object is pixel art you draw yourself: a 16x16 grid per object, every cell a hand-placed colour, rendered as a 3D block (a darker extrusion down-right, a thin bevel on every cell) with a soft contact shadow. The silhouettes are lit like a heat camera: the heat comes from how thick the body is at each point, so the nose, the hair and the fingertips run cooler, plus one hotspot and a slow moving noise, mapped through a black > red > orange > cream ramp. The edge stays razor sharp, with a soft glow outside it. Motion rules: nothing pops in (eased glides of about 0.5s with a soft start), fast moves get real motion blur instead of ghost copies, and captions type with the voice behind a block cursor. Banned: stock footage, AI video, 3D camera moves, particles, glitch. </direction> <structure> Hang every scene on the transcript's word timings (measure them with faster-whisper word timestamps, one short chunk at a time): 1. The first word fills a yellow card. The card wipes off to the left and the question types underneath on two centred lines that glide to stay centred as they grow, the key word underlined in red as it's spoken. 2. The objects streak past in lanes above and below the words, a cyan four-point flare sweeps in, then the page burns away: black closes in from the edges onto the head, behind a hot red edge. 3. The head heats up on black while the answer types low on the left. A thin white orbit draws itself around the head, its far half passing behind it. On the last word, paper blooms out from inside the head with the objects in it, and the camera falls through: the head rushes past the lens until the paper is all we see. 4. The objects sit on a tilted ring seen in perspective (the front ones bigger and lower, drawn last), slowly turning, each one bobbing; the coin flips and lands with sparks. 5. On each of the 3 words, that object flies out of the ring to the left and grows, the page floods black from it, and the word types beside it (camera: a red shutter star; book: a red ribbon sweep; record: two white orbits and red notes). Then the flood drains back into the object as it flies home, and the camera dives closer into the ring. Put each hero object in the slot that is front-left when its word lands. 6. The ring pulls in a touch, bursts outward, every object turns to ink, and ink blots grow and swallow the page. 7. The hand rises into frame, hot, with the line split either side of it. The heart drops into its palm, then lifts to the centre at the size the next scene holds it. 8. The last word, one letter per beat, the background flipping paper / black / red, the heart holding its place across the cut, the centre object morphing pixel by pixel into the next one. 9. Thin brush strokes sweep through, the letters drift apart and float, then settle back into the word with the heart above it. The camera keeps pushing through the hold. </structure> <build> 1. One HTML canvas. Every frame is a pure function of time inside seek(t). 2. Bake each pixel sprite once to an offscreen canvas (the extrusion layers, then the face), then draw it scaled and rotated. 3. The heat: precompute each silhouette's base heat once (a distance transform of the mask, held at the edge value outside it so nothing dark bleeds in when it is upscaled). Every frame, add the hotspot and two octaves of value noise at quarter resolution, map it through a 256-step lookup, upscale it, and cut it with the full-resolution mask. 4. The fall through the head: draw the ring scene into an offscreen layer, mask it with the head scaled up about a point inside the skull, and grow that scale on a cubed curve until the head covers the frame. The ring inside zooms less than the head does, so it reads as depth. 5. The pixel morph: pair every cell of object A with a cell of object B and glide each one's position and colour, with a small random delay per cell. 6. Render with Playwright at 60fps with 8 motion-blur subframes, then lay the original audio back on with ffmpeg. </build> <gotchas> A mask PNG with no alpha channel fills the whole box with heat, so put the silhouette in the alpha. A limb cut by the photo's border shows as a straight edge, so extend the mask past the frame. A handoff only reads as smooth when the last frame of one scene is the first frame of the next: draw the outgoing scene under the transition, and give scenes that show the same ring one shared zoom function. A flood that's still retreating when the next scene starts reads as a hard cut, so let it finish first. A fast ease-out flood reads as a jump; use a soft start. </gotchas> <start> Ask me for the inputs, measure the word timings, then show me 8 stills before you render. </start>