MiniMax H3 Prompt Guide: Directing a 15-Second 13-Cut Fashion Film with 5 Looks

Learn how to direct a 15-second fashion editorial video in MiniMax H3, preserving character consistency across 13 rapid cuts, 5 styling looks, and kinetic typog

tau · October 3, 2026

#MiniMax H3 #AIVideo #PromptEngineering #FashionFilm #CharacterConsistency

MiniMax H3 Prompt Guide: Directing a 15-Second 13-Cut Fashion Film with 5 Looks

AI visual creator Ima Studio (@ImaStudio_ai) has shared an advanced prompt structure for MiniMax H3, demonstrating how to create a 15-second 'Rose Glow' fashion magazine trailer featuring 13 rapid cut transitions and 5 distinct outfits while locking character identity to a single reference image.

Frame from a 15-second 13-cut Rose Glow fashion film and lookbook concept video generated using the MiniMax H3 model

Image source: Ima Studio (@ImaStudio_ai)

MiniMax H3 is a general-purpose omni-modal generation model capable of processing unified text, image, and audio contexts to generate up to 15 seconds of 2K video accompanied by native stereo sound. This practical guide breaks down the structured prompt blueprint designed to eliminate facial drift, balance live-action cinematography with motion graphics, and assemble a magazine cover sequence.

Core Architectural Blueprint and 5-Look Studio Staging

Rather than relying on vague descriptive prose, Ima Studio's prompt modularizes video parameters, identity constraints, spatial boundaries, outfit definitions, cut-by-cut sequencing, and live-action/graphics ratios to maximize prompt adherence in MiniMax H3.

  • Baseline Specifications: 15 seconds duration, 16:9 aspect ratio, 24fps live-action fashion magazine trailer format.
  • Art Direction: Korean/Asian luxury fashion editorial × pink campaign aesthetic × kinetic typography.
  • Identity Lock: Anchors the reference image as the sole identity source, maintaining the same young adult East Asian woman, facial structure, skin tone, bangs, long dark-brown hair, and body proportions across every cut.
  • Spatial Containment: Strictly confined to an indoor pink fashion magazine studio to prevent background leakage into outdoor, café, or generic settings.
  • Composition Ratio: 55% live-action cinematography balanced with 45% motion graphics integration.

To inject visual dynamism into a short runtime, the prompt explicitly predefines five coordinated outfits:

  1. LOOK 01: Pink tweed jacket + mini skirt
  2. LOOK 02: Cream-white top + pink-and-white asymmetric skirt
  3. LOOK 03: Rose fitted mini dress
  4. LOOK 04: Pale-pink Y2K crop top + pink-gray low-rise skirt
  5. LOOK 05: Pale-pink satin/tweed hero dress

The studio set is anchored around soft rose-pink and cream-white tones with black/silver accents, introducing varied design elements—magazine pages, typography walls, acrylic panels, mirrors, torn paper textures, grids, flowers, contact sheets, and checkerboard patterns—across individual cuts.

13-Cut Sequence and Kinetic Typography Breakdown

Over the 15-second timeline, 13 cuts synchronize model posture, camera tracking, and kinetic typography overlays.

Cut #Look / SettingCamera Movement & Model ActionTypography & Motion Graphics
01Opening close-upSide gaze shifting directly toward cameraROSE typography reveal
02LOOK 01 setGentle face touch with camera push-inGLOW text integration
03Cream beauty setBeauty close-up brushing hairCHINA MUSE text
04LOOK 02 setSide-to-front turn within a pink circle graphicSOFT ATTITUDE text
05Giant magazine pageHalf-step forward with editorial posingTHE PINK EDIT text
06LOOK 03 setSeated posture on a silver platformMODERN ROMANCE text
07Chest-up shotModel adjusting an earringCOVER GIRL text
08LOOK 04 / checkerboardHip shift with a look back over the shoulderNEW MOOD text
09Centered model compositionCentered framing framed by 3–4 photo contact sheetsContact sheet layout graphics
10Torn pink paper revealShoulder turn revealed through torn paper textureFASHION ISSUE reveal
11LOOK 05 setStepping forward between pink acrylic and silver framesROSE GLOW text
12Pull-back shotCamera pulls back as magazine pages and frames assembleSOFT ATTITUDE text
13Final cover assemblySlight turn into direct gaze; hold final 0.5sGraphics assemble into ROSE GLOW / CHINA MUSE / SOFT ATTITUDE / VOL. 2026 / THE PINK EDIT

Complete MiniMax H3 Prompt Template

Copy the prompt below directly into the MiniMax H3 (or Hailuo AI) video generation prompt box and attach your character reference image as the primary input:

Create a 15s, 16:9, 24fps live-action fashion magazine trailer with exactly 13 cuts. Style: Korean/Asian luxury fashion editorial × pink campaign × kinetic typography. Theme: ROSE GLOW. Tagline: SOFT ATTITUDE.
Use the reference girl as the ONLY identity source. Keep the same young adult East Asian woman, face, skin tone, age, long dark-brown hair, bangs, slim body and proportions. No face drift, extra people, hairstyle or body changes.
WORLD: PINK FASHION MAGAZINE STUDIO only. No outdoor/city/café/hotel/nature. Every cut changes set, layout and composition. Palette: soft/rose pink, cream white, small black/silver. Use magazine pages, typography walls, acrylic, mirrors, torn paper, grids, flowers, contact sheets, frames, checkerboard.
LOOKS: 01 pink tweed + mini skirt; 02 cream-white top + pink-white asymmetric skirt; 03 rose fitted mini dress; 04 pale-pink Y2K crop top + pink-gray low-rise skirt; 05 pale-pink satin/tweed hero dress.
CUTS: 1 close-up side gaze→camera, “ROSE”. 2 LOOK01, touch face, push-in, “GLOW”. 3 beauty close-up, brush hair, cream set, “CHINA MUSE”. 4 LOOK02, side→front, pink circle, “SOFT ATTITUDE”. 5 giant magazine page, half-step, “THE PINK EDIT”. 6 LOOK03 seated on silver platform, “MODERN ROMANCE”. 7 chest-up, adjust earring, “COVER GIRL”. 8 LOOK04, checkerboard, hip shift + look back, “NEW MOOD”. 9 centered model + 3–4 photo contact sheets. 10 torn pink paper reveal, shoulder turn, “FASHION ISSUE”. 11 LOOK05, pink acrylic + silver frame, step forward, “ROSE GLOW”. 12 pull-back as magazine pages/contact sheets/frames build around her, “SOFT ATTITUDE”. 13 final cover: slight turn→direct gaze; graphics assemble into ROSE GLOW / CHINA MUSE / SOFT ATTITUDE / VOL. 2026 / THE PINK EDIT. Hold final 0.5s.
55% live-action / 45% motion graphics. Each cut combines character action + camera move + typography + graphics. No dancing. Avoid slideshow, repeated backgrounds, sticker/PPT feel, cyberpunk, unreadable text, logos, watermark, UI, subtitles.

Negative Constraints and Pacing Controls

To maintain high visual fidelity and prevent unwanted generation artifacts, keep the following constraints in mind:

  • Eliminate Dance Motions: Restricts movement to subtle fashion poses and avoids erratic dance choreographies.
  • Prevent Static Slideshow Artifacts: Pairs each cut with specific camera motions and kinetic typography to avoid rigid PowerPoint-style transitions.
  • Suppress UI and Synthetic Watermarks: Disallows overlayed subtitles, logos, or user interface elements to preserve a clean, professional finish.
  • Reference Image Selection: Choose a high-clarity portrait with distinct facial features and hair contours to ensure optimal identity lock across cut transitions.

Original source