This clip was generated with MiniMax H3 by MiniMax. The prompt on this page is the exact text used to create it.
The example was generated with MiniMax H3 using: 16:9 · 0:15 · Anime · Fantasy · Cinematic. Reusing these settings gives the closest match.
Copy the prompt as-is first, then change the subject while keeping the camera, motion, and lighting terms. Video models are sensitive to wording, so expect a few runs.
Yes. Prompts work best as starting points — swap scenes, adjust pacing terms, or change the aspect ratio to fit your project.
High-quality anime footage. The world setting is a refined modern fantasy academy, depicted as 2D hand-drawn slice-of-life conversation animation. Facial expression design must always be done in this order: eyes → eyebrows → eyelids → mouth → gaze. [Reference Images and Character Consistency] Image 1 is the reference image for Character A. Maintain the character depicted in Image 1 as the same person throughout the entire video. Use Image 1 only as a reference for Character A's face, eye shape, iris color, hairstyle, hair color, outfit, accessories, physique, silhouette, atmosphere, and character colors. Image 2 is the reference image for Character B. Maintain the character depicted in Image 2 as the same person throughout the entire video. Use Image 2 only as a reference for Character B's face, eye shape, iris color, hairstyle, hair color, outfit, accessories, physique, silhouette, atmosphere, and character colors. Do not reproduce the background, room, furniture, pose, camera angle, framing, or lighting arrangement of the reference images themselves. Only the following may change: facial expression, gaze, mouth shape, posture, breathing, and natural movement of hair and clothing. Do not blend the features of Character A and Character B. Averaging faces, swapping outfits, swapping hairstyles, turning them into different people, cloning, creating doppelgangers, and adding extra people are prohibited. Even if Image 2 contains a weapon or handheld item, Character B's hands must hold nothing. Only if Image 1 contains a non-combat accessory or small bag may it remain naturally on Character A. However, the bag must be fixed to the body side; holding it in the hand, opening it, or using it in action is prohibited. No weapons, large equipment, or glowing combat gear. [Props, Color Tone, and Atmosphere Adjustments] The location, seating position, conversation content, the action of dividing the bread, and the camera composition are fixed. On that basis, naturally adjust hair ornaments, brooches, ribbons, bags, and other accessories, props, color schemes, light impression, and background accent colors to match the reference images of Character A and Character B. Adjust the overall line color, shadow color, background accent colors, and color temperature of the afternoon light to match the character colors and atmosphere of Character A and Character B. However, maintain the bright, refined, and clean everyday feel of a modern fantasy academy; do not make it a dark combat space, horror, decadent, or extremely backlit. [Art Style Fixation] Thin, delicate colored line art; transparent two-to-three-layer colored shadows; soft gradients. Do not make saturation too low; naturally reflect the reference characters' inherent colors into the hair, eyes, clothing, reflected light, and background accent colors. Eyes have jewel-like iris depth and multiple small catchlights; hair has pearl-like multi-layered highlights. Draw hair, eyelashes, lace, embroidery, hair ornaments, accessories, and clothing wrinkles with high density. Fabric is rendered with distinct delicate textures for high-quality matte material, lace, metal, gemstones, and ribbons. The background uses the same 2D hand-drawn brushwork, drawing the cream-colored walls, wooden bench, plants, and distant school building with sufficient density. Thick lines, heavy outlines, simplified TV anime style, flat single-layer cel shading, low-density backgrounds, smooth 3D assets, semi-realism, photorealistic textures, heavy dulling, AI-average faces, and mixed art styles are prohibited. [Fixed Location and Character Placement] A wooden bench in the school courtyard after class. Character B is on the left side of the screen, Character A on the right. The two sit next to each other. Only Character A holds one round anpan (sweet bean bun) wrapped in thin paper, supporting it from underneath with the right hand. At the start, the bun is perfectly round with no bite marks or missing pieces. The fingers are placed under the paper wrapper, away from the eating position. Do not multiply the bun, the paper wrapper, or the fingers. [Cut 1] Camera position: front-center of the bench. Camera height: at the eye level of the two seated characters. Frame both characters from the front, from the chest up, at equal size, with the background vanishing axis receding between the two into the distance. Purpose: to establish the relationship between the two and that the bun belongs to Character A. Character B looks at the bun, then returns their gaze to Character A, brings both hands together at the chest, and leans slightly closer with the whole body. Eyes sparkling, eyebrows raised, eyelids opened in anticipation, making a coaxing mouth shape, gaze directed at Character A. Only Character B speaks in a bright, sweet voice: "Let me have a bite of that anpan." A brief pause, then coaxing further: "Please, it's the request of a lifetime!" Character A does not speak, but looks at Character B with an exasperated yet gentle expression. [Cut 2] Camera position: diagonally in front of Character A. Camera height: at the eye level of the two seated characters. Angle from over Character A's shoulder toward Character B, showing both faces and the bun in a medium close-up. The background vanishing axis shifts to the upper right. Purpose: to show the response, the offering hand, and the single bite action in one continuous sequence. Character A softens the eyes, lowers the eyebrows slightly, gently narrows the eyelids, forms a wry smile with the mouth, and directs the gaze at Character B. A small exhale, a brief pause, and only Character A speaks in a calm, gentle voice: "Honestly, you're hopeless. Just one bite, okay?" While finishing the line, Character A offers the single anpan in the right hand toward Character B's mouth. The owner remains Character A; it is not handed over. Before contact: the bun is perfectly round. There is a small gap between Character B's teeth and the bun. Contact: show the moment Character B leans in and the teeth touch the upper edge of the bun, taking just one bite. After contact: only a small crescent-shaped bite is missing from the bun, and Character B pulls the face away. Character A returns the same bun to the chest with the right hand. Do not bite the hand or fingers. Do not duplicate the bun, paper wrapper, or fingers. [Cut 3] Camera position: diagonally in front of Character B. Camera height: at the eye level of the two seated characters. Angle from Character B's side toward Character A, showing Character B slightly larger from the chest up, then briefly panning left to a chest-up two-shot of both. The background vanishing axis shifts to the upper left. Purpose: to place Character B's satisfied face and Character A's returning smile in the same final frame. Character B closes the mouth and chews twice, small bites, then swallows. After a brief pause, the eyes brighten, the eyebrows form a soft arc, the eyelids narrow happily, a satisfied smile forms on the mouth, and the gaze returns to Character A. Only Character B speaks in a buoyant, gentle voice: "Mmm, delicious!" Character A holds the same bun with one bite missing at the chest with the right hand, silently returning a warm smile. Keep both faces and the bun in the same frame, and cut at the moment Character B finishes the final vowel of "delicious" and both smiles are complete. [Audio] Generate two Japanese voices that match the characters' appearance, perceived age, and atmosphere. Character B has a bright, sweet voice; Character A has a calm, gentle voice. Do not swap speakers; deliver the three lines verbatim, without overlapping, and synchronize mouth movements with the vocalization. No dialogue other than the lines, no narration, no subtitles, no on-screen text. No BGM, no music. Only ambient sound and foley. Synchronize weak wind, distant birds, fabric rustling, the paper sound of the wrapper, and the small compression sound at the moment of the single bite into the soft bread with the visible actions. Do not add exaggerated chewing sounds. [Important Prohibitions] No speaker swapping, no simultaneous speech, no lip-sync errors, no duplication of the bun, no unsolicited handover, no eating more than one bite, no bite marks before eating, no deformation of fingers, no adding people, no foreground covering faces, no light spots in the center of faces, no subtitles, no logos.