How to Make Sumi-e AI Videos with Pollo.ai and MiniMax H3
How to Make Sumi-e AI Videos with Pollo.ai and MiniMax H3
Sumi-e (墨絵) is the old Japanese art of ink wash painting. It uses black ink, soft brush strokes, and empty space to create calm, powerful images. Today, you can bring this beautiful style to life as video, using AI tools.
In this guide, I will show you exactly how I created a Sumi-e style AI video. I used Pollo.ai as my main platform, and the video itself was generated with MiniMax H3. You do not need any drawing skill or video editing experience. You just need the right prompts, and I will give them to you.
Why I Used Pollo.ai for This Project
Pollo.ai is one of the most popular all-in-one AI video platforms in 2026. Instead of using one single AI model, Pollo.ai gives you access to many top video models in one place. This includes models like Kling, Runway, Veo, Hailuo, and MiniMax, all inside one workspace.
This is why Pollo.ai works so well for creators. You do not need five different subscriptions. You just log in to one platform and pick the best model for your project. For this Sumi-e video, I picked MiniMax H3 inside Pollo.ai because it gave me the most natural ink motion and detail.
Pollo.ai also supports text-to-video, image-to-video, and video-to-video generation. It has a free tier for beginners, and paid plans for creators who want higher resolution and more monthly credits. This makes it a good starting point if you are new to AI video and want to test different models before choosing one.
The Tool Behind the Video: MiniMax H3
The actual video in this tutorial was generated using MiniMax H3, a new AI video model. MiniMax H3 is different from older video models because it treats text, images, video, and audio as one connected system. This means it can generate video and matching sound together, in a single generation.
MiniMax H3 supports native 2K video at 24fps, with clips from about 5 to 15 seconds, and it can be extended further. It also accepts up to 9 reference images in one prompt. This is the exact feature we will use in Step 2 below, where we combine a storyboard image with a personal photo to create the final Sumi-e video.
Because of its strong image understanding, MiniMax H3 is a great choice for artistic, style-heavy videos like Sumi-e ink wash animation. It keeps character details, ink texture, and motion consistent across the whole clip.
Step 1: Create the Storyboard with GPT Image 2
Before making the video, we need a storyboard image. This storyboard sets the visual style, mood, and layout for the whole project. For this, we use GPT Image 2. (Prompts are below).
Step 2: Turn the Storyboard into a Video with MiniMax H3
Now it’s time to animate everything. Open MiniMax H3 inside Pollo.ai, and use its reference mode, which allows more than one input image.
You will upload two images:
- Your storyboard image from Step 1 (this sets the Sumi-e ink style, layout, and mood)
- Your own picture (this becomes the subject or character inside the video) then you will add the below prompt.
GPT Image 2 Prompt:
Create a premium minimalist cinematic storyboard page for an experimental hiphop ink-motion short film titled “Esso AI Show” The layout should feel like a professional director’s previs board mixed with luxury fashion motion design and Japanese ink calligraphy aesthetics. White clean background with ultra-minimal graphic design. Use only black ink, grayscale tones, and selective deep crimson-red accents. No additional colors. The storyboard contains 8 sequential action panels arranged in a clean horizontal grid layout with large negative space and elegant spacing. Each panel shows a simplified mannequin-like human figure performing dynamic hiphop-inspired brush choreography using a giant calligraphy brush. The brush creates explosive ink arcs, circular ink rings, sweeping strokes, liquid splashes, and flowing abstract motion trails across the frame. Ink should feel energetic, painterly, and physically fluid — combining sumi-e brushwork, cinematic motion blur, and fashion-editorial abstraction. Visual language: Japanese ink wash minimalism abstract motion design cinematic storyboard presentation IMAX-scale movement energy dynamic wide-angle motion arcs expressive negative space clean modern typography luxury monochrome graphic identity rough dry-brush textures mixed with dense wet ink high contrast rendering elegant directional arrows and timing labels minimal but readable production notes Panel progression: Initial upward ink swing creating curved arc Forward brush thrust with long ink trail Massive rising ink wave engulfing the figure Circular spinning motion forming ink ring Aggressive forward pressure blast Full-body turning motion with spiral ink wrap Extreme tension pullback before release Final explosive ink impact flooding frame outward Typography style: modern minimalist editorial typography tiny cinematic production labels shot timing indicators subtle arrows and movement directions hand-painted logo text “Inkflow” clean Swiss-style layout structure Composition: large white margins balanced panel spacing asymmetrical elegance cinematic readability storyboard presentation board aesthetic motion study sheet feel professional art-director presentation quality Style references: Japanese sumi-e ink painting + luxury fashion storyboard + cinematic previs art + abstract dance motion graphics + modern IMAX concept art. Negative prompt: No realistic faces. No detailed anatomy. No colorful background. No photorealism. No comic-book style. No anime style. No cluttered text. No messy composition. No extra props. No unnecessary environment. No 3D rendering look. No low-resolution textures. No watermark.
MiniMax H3 Image to Video Prompt:
@Image1 = storyboard structure and motion rhythm reference
Create a premium IMAX-scale cinematic ink-dance sequence inside an infinite seamless white void with no visible floor horizon. The visual language should feel like a fusion of avant-garde hiphop choreography, luxury fashion film, abstract Japanese ink painting, and large-format IMAX camera energy. Every movement must feel physically grounded yet mythic, with aggressive cinematic momentum, dramatic negative space, and extreme wide-lens perspective distortion. Main subject: a hiphop dancer wearing a loose oversized white top and dark oversized pants with irregular black ink erosion along fabric edges. The costume behaves organically, absorbing and releasing ink through movement. The dancer holds a massive heavy ink brush in the right hand at all times. The brush is never static — every swing, drag, rotation, stop, and acceleration creates dense wet ink trails, suspended ink particles, calligraphic streaks, and volumetric ink masses in the air. Visual style: ultra high-contrast monochrome minimalism with pure white background and only deep black ink, grayscale gradients, and controlled crimson-red accent ink. No other colors. Strong IMAX visual clarity, cinematic motion blur, soft film grain, liquid ink physics, dramatic negative space, luxury fashion editorial framing, elegant high-speed cinematography, dynamic lens breathing, subtle shutter smear, large-scale atmospheric depth. Camera language: aggressive IMAX handheld precision mixed with floating crane movement, low-angle tracking, ultra-wide 24mm lens distortion, kinetic push-ins, rotational drift, proximity lens passes, dynamic speed ramps, and physically immersive motion. Camera should feel reactive to the dancer’s momentum, almost fighting against the energy of the ink explosions. Sound design and music: Minimal hiphop beat with deep groove and cinematic spatial bass. 0:00–0:08 → steady groove with clean kick rhythm and subtle sub bass. 0:08–0:11 → increased rhythmic density and rising tension. 0:11–0:13 → near silence with suspended atmospheric pressure. 0:13–0:15 → massive bass impact synchronized with final ink strike. TIMELINE 0:00–0:02 Low-angle 24mm tracking close shot. The dancer shifts weight sharply in rhythmic groove motion while stepping diagonally toward camera. The brush whips from behind the body, generating the first long liquid ink arc across empty white space. Ink subtly bleeds along clothing seams and sleeves. Camera glides beside the movement with slight handheld vibration. SFX: soft brush drag, fabric movement, low bass pulse. 0:02–0:04 The dancer accelerates into rapid footwork while rotating shoulders aggressively. The brush carves horizontal ink ribbons through the air. Camera circles tightly around the body with dynamic parallax distortion. Ink trails linger volumetrically before dissolving into smoke-like particles. SFX: rhythmic foot friction, layered brush slicing, distant sub resonance. 0:04–0:06 Camera violently pulls backward then crashes forward again. The dancer drops low and explodes upward into a vertical rotational swing. Massive black ink erupts upward like liquid architecture. Clothing expands dramatically from momentum. Wide-angle distortion exaggerates body rotation and brush velocity. SFX: rising air pressure, heavy ink whip, bass pulse expansion. 0:06–0:08 Slow rotational drift around the dancer while continuous spinning footwork generates circular calligraphic ink structures wrapping around the body. Arms extend wide as ink spirals interweave with silhouette and fabric motion. The dancer momentarily disappears behind flowing ink mass before re-emerging. SFX: circular brush resonance, layered fabric movement, hypnotic ambience. 0:08–0:10 Camera retreats rapidly as if pushed away by force. The dancer advances aggressively toward lens with increasingly violent brush swings. Ink density becomes heavier and more chaotic, creating giant suspended arcs that nearly consume the frame. Subtle crimson-red ink accents briefly emerge within black ink collisions. SFX: intensified groove rhythm, low-frequency vibration, pressure-like ambience. 0:10–0:12 Sudden aggressive push-in toward the dancer. Using forward rotational momentum, the dancer performs a massive turning motion while the brush generates a full circular ink ring around the body. The figure passes extremely close to lens edge, briefly disintegrating into fragmented ink strokes before reconstructing physically. SFX: distorted bass hit, rapid brush turbulence, cinematic suction effect. 0:12–0:13.5 Near silence. Time feels suspended. The dancer slowly pulls the heavy brush far behind the body into extreme tension. A long dense ink tail stretches across the void while surrounding floating ink retracts inward unnaturally, as if gravity reverses. Camera slowly pushes closer into the tension. SFX: minimal sub bass pulse, stretched brush fibers, soft atmospheric air. 0:13.5–0:15 Full-force final impact. The dancer lunges explosively toward camera, violently whipping the brush directly into lens. Massive calligraphic ink strokes flood the entire frame with overwhelming force and liquid realism. Black ink completely consumes the white environment. The HelmyPad Film logo emerges organically inside the spreading ink texture on a fully black painted background — elegant, cinematic, premium, not graphic-design-like. Final frame should feel iconic, explosive, and theatrical. SFX: massive bass impact, violent ink explosion, deep cinematic resonance. Negative instructions: No cartoon look. No anime style. No cheap VFX look. No low-detail ink simulation. No static camera. No flat lighting. No extra characters. No colored lighting. No visible studio environment. No text except final AIToolsGuide Film logo reveal. No glitch effects. No artificial AI morphing artifacts. No floating disconnected limbs. No unrealistic body deformation. No low-energy choreography. No repetitive movement loops. No oversaturated reds. No muddy grayscale.
