After watching Spider-Man: Brand New Day, I kept noticing how distinct each major scene felt from the last. That observation turned into a small project: instead of generating a single static illustration, I wanted to produce three separate movie-like moments from one character.
I set out to build three scenes with clearly different moods: a quiet rooftop shot, a high-motion swing through rain, and a hopeful sunrise moment. I ran this in PixAI Studio, building an AI cinematic scene generator workflow with Tsubaki.3 as the generation model throughout.
The approach was simple: three Image Nodes in a single workspace, one shared Spider-Man character direction, and a different scene direction pushed into each node. Same character, same model, three distinct outputs.
Three completely different Spider-Man scenes, created with the same character direction and model
Below is the exact setup, scene by scene, with the prompts included.
Set Up a Simple Cinematic Scene Workflow in PixAI Studio
I started in PixAI Studio, using its node-based canvas to structure an AI movie scene generator workflow. Rather than generating one image, reviewing it, then starting a fresh generation, I kept all three scene experiments visible in the same workspace at once.
I added three Image Nodes to the canvas, one per scene. Each node ran on Tsubaki.3, which kept the character consistent across nodes while letting me vary only the scene direction.
Three Image Nodes side-by-side in one workspace, each using Tsubaki.3
With the workspace configured, the next step was defining a single base character direction to reuse across all three nodes.
Start with One Basic Spider-Man Direction
I kept one consistent core description across all three prompts: Spider-Man in the classic red-and-blue suit, rendered as a cinematic movie still. That single anchor gave every generation the same starting point, while still leaving room to push each scene toward a different result.
From there, only five variables changed per node: scene, action, camera angle, lighting, and atmosphere. Node one placed Spider-Man on a rooftop at night. Node two turned that into a dynamic web-swinging shot in heavy rain. Node three moved to a calmer rooftop setting at sunrise.
Isolating those variables made the comparison meaningful. Instead of running the same composition three times, I could directly observe how changing scene, angle, lighting, and atmosphere shifted the tone: from dark and watchful, to energetic and rain-soaked, to bright and resolved.
Scene 1: Spider-Man on a Rooftop at Night
The first scene was built to be quiet: no action, no motion blur, just a single still moment before the night starts. It works as an opening entry in a superhero cinematic scene series because it establishes the character before anything happens.
Prompt used:
"Spider-Man crouching on the edge of a New York City rooftop at night, overlooking the glowing skyline, dramatic cinematic composition, medium-wide shot, strong silhouette, blue and red suit illuminated by city lights, deep shadows, atmospheric night sky, cinematic movie still"
Spider-Man crouching over Manhattan, waiting for the night to start
He's crouched right at the rooftop edge, half in shadow, with the skyline carrying most of the visual weight behind him. The medium-wide framing was a deliberate choice: it gives the city enough space to read as massive next to him, reinforcing the quiet-before-the-action tone the prompt was built for.
Scene 2: Spider-Man Swinging Through a Rainy City
The second scene was built as the direct inverse of Scene 1: motion, weather, and chaos, all layered into one frame. This is the kind of output that demonstrates what makes Spider-Man AI art interesting to iterate on: high energy, unconventional angles, and a city environment that reads as alive.
Prompt used:
"Spider-Man swinging through a rainy New York City at night, web line stretched overhead, skyscrapers surrounding him, heavy rain, wet streets reflecting city lights, dramatic low-angle perspective, dynamic action composition, cinematic movie still"
Mid-swing, rain hammering the street below
The low-angle perspective is doing most of the work in this frame. Looking upward at him mid-swing, with rain streaking through the composition and the illuminated city below, produces a sense of velocity that the rooftop scene never had. Same character, same suit, an entirely different energy profile.
Scene 3: A "Brand New Day" Sunrise Scene
The third scene closes on a hopeful note, matching the film's own title directly. After the stillness of Scene 1 and the motion of Scene 2, this one needed to function as resolution: a new day, rendered literally.
Prompt used:
"Spider-Man standing on a rooftop at sunrise, looking out over New York City as the first golden light fills the skyline, warm sunlight breaking through the clouds, hopeful cinematic atmosphere, dramatic wide composition, red and blue suit glowing in the morning light, cinematic movie still"
Same city, new light. Spider-Man watching the sunrise
Switching to a standing pose instead of crouching or swinging made the largest difference in this scene. He's upright, still, and facing the light rather than reacting to anything in frame. Combined with the warm gold tones across the skyline, the result reads like exactly what you'd want out of an AI movie still generator: one frame that tells a complete story without needing the other two scenes for context.
The direction he's facing turned out to matter more than expected. Rather than looking back toward the camera or down at the street, he's turned outward, toward the horizon. It's a small compositional detail, but it's what keeps the image from reading as a posed portrait and instead makes it feel like a single frame lifted from a larger sequence, which is the exact closing note I wanted after the tension of Scene 2.
Compare the Three Cinematic Directions
With all three scenes generated, I lined them up side by side to check how far one character could stretch across different moods using the same base direction.
Same character, three moods side-by-side: quiet, dynamic, hopeful
Composition shifted the most between the three. Scene 1 used a medium-wide shot that gave the character and the skyline roughly equal visual weight. Scene 2 switched to a low-angle, action-first composition centered on mid-motion. Scene 3 went wide again but shifted to an upright, static pose facing outward.
Lighting carried a lot of the emotional signal too. Cool blues and city glow define Scene 1. Heavy grays and wet-street reflections define Scene 2. Warm gold tones define Scene 3. None of the three prompts specified emotion directly. The lighting variable alone did most of that work.
Out of the three, Scene 2 reads the most like an actual movie still to me. The combination of low-angle perspective, rain, mid-action swinging, and an illuminated city produces the strongest sense of a frame pulled from a longer sequence. Scene 1 functions more like an establishing shot. Scene 3 plays more like a closing beat.
The suit itself is a useful case study in how much a fixed design element can be reshaped by lighting alone. In Scene 1, the red and blue read muted and shadow-heavy, blending into the night skyline. In Scene 2, those same colors turn saturated and reflective, picking up the wet-street glow. In Scene 3, they shift toward gold, catching direct sunrise light. None of the three prompts described the suit differently. Only the light source changed, which says a lot about how much mood in an AI cinematic scene generator workflow comes from lighting rather than character design.
That's really the core mechanic behind cinematic AI art like this: the character stays nearly constant, but a different setting, angle, and light source is enough to turn one hero into three distinct stories.
How to Create More Spider-Man Cinematic Scenes
This setup generalizes well beyond these three scenes. Once a base character direction is locked in, the fastest way to produce a new mood is to adjust setting, action, camera angle, lighting, or atmosphere while keeping the core character direction fixed.
A few directions worth testing: a quiet subway platform, a nighttime alley, a bridge scene, snowy New York, a city street at dusk, or a daytime swing through the city. Each only requires a new scene prompt, and the rest of the workflow stays identical.
I didn't generate every option above for this piece. They're starting points if you want to keep experimenting inside the same PixAI Studio workspace. If you haven't used Studio yet, this introduction to PixAI Studio covers the node-based canvas basics before diving in.
One practical way to approach this: treat each new prompt as an isolated experiment. Keep the core character direction fixed, then change one or two visual decisions per iteration. A low-angle camera reads as more imposing. A wide shot gives the city more visual weight. Cool nighttime lighting pushes toward tension, while warm sunrise lighting pushes toward hope, even with the same character. That's what makes an AI cinematic scene generator workflow more useful than generating a new picture at random each time. You're directing the frame deliberately instead of leaving every visual decision to chance.
I Shared My Favorite Cinematic Scene on PixAI
Once all three scenes were finished, I published my favorite directly from PixAI Studio. Scene 2, the rainy web-swinging shot, was the one I ended up sharing. It's the most dynamic of the three, and it holds up as a standalone piece rather than only working inside a comparison. The rain, the low-angle perspective, and the wet-street reflections combine to give it a stronger sense of motion than either of the quieter rooftop scenes.
Published on PixAI: Spider-Man: Rainy Night Swing
You can view the full published piece here.
Conclusion
Three Image Nodes, one base character, Tsubaki.3, and three distinct Spider-Man moments. That's the entire mechanic behind this workflow: change the scene, camera angle, lighting, and atmosphere, and the same character can carry a completely different feeling each time.
Watching Brand New Day is what pushed me toward this experiment in the first place, and building these three scenes was a good way to extend that feeling a little longer. If you want to try a similar setup yourself, PixAI Studio is a reasonable place to start.
Top comments (0)