“Klaus” (Academy Award-nominated) used AI to infer consistent 3D surface information from hand-drawn 2D animation. As Sergio Pablos explains, the key problem is surface correspondence. The system must track the same point on the character across different drawings. This creates something similar to a persistent UV map over 2D animation. Once that exists, textures, lighting and other surface information can follow the character’s volume without using a fully animated 3D model. AI could reconstruct fake 3D passes directly from 2D drawings: • normals • depth • position • material masks • UV / surface correspondence Those passes could drive lighting, shadows, materials and FX while preserving the original 2D animation. Pipeline: ✦ Original 2D animation ⬇ AI-generated textures ⬇ Reconstructed 3D passes ⬇ Lighting / shading / FX It is almost the inverse of Arcane: 3D → made to feel 2D Instead: 2D → given the depth and material richness of 3D What concerns me most is that artists who are open to AI-assisted workflows can face harassment or backlash, making some afraid to collaborate even when they are interested. I really hope this changes over time. I'm particularly interested in workflows involving: • color transfer • rendering and relighting • illustrated texture transfer • AI-assisted compositing • hybrid shots combining hand-drawn animation and generated elements • sequences mixing traditional / handcrafted shots, • hybrid shots mixed fully generative shots in final animation This is the link for the full interview with Sergio Pablos: dailymotion.com/video/x7rcg1…

With Wan 3.0 AI anime is getting seriously close to TV quality. You can create sequences like this one that look like an actual seasonal anime. I made a full fantasy fight between a Mage and a Swordswoman — and it genuinely feels anime-ready. Seedance 2.5 keeps characters more on-model. Wan is better at timing, poses and complex layouts. Hailuo H3 sits in between. All three support Vid2Vid and Omni Reference, so you can pass shots and references between them. The possibilities are endless. Interesting times.

GPT-6 Astra + 3D Jutsu I tested this workflow, and I am actually impressed by Astra's understanding of camera work. It seems useful for certain scenes where you need to lock positions and camera movements very accurately. The way to work with this is: 1. Use basic shapes: prisms, spheres and cylinders. 2. No details. 3. Do not add animations; only animate their placement in the scene. 4. Give each subject a different solid color. Rendered in Seedance 2.5 at 480p. When you add detailed elements like animations or geometry that accurately matches the subject, you don't give the model room to imagine, and it just renders directly on top of the scene's geometry.

Since I noticed in my last post that many people love dragon art, I decided to make a dragon animation... (I summoned one)

I finally managed to create an Anime Fight Scene that looks straight out of a real TV show! The latest AI updates and my recent research into animation pipelines are paying off massively, opening up entirely new techniques to explore. Bringing this together required multiple layers and processes, but I’m thrilled with the final result.

If you don’t know what "LATENT SPACE" is, you’re missing one of the most important concepts behind AI Video. Fortunately, Dr. Kit explains it in just 1 minute. Watch this 👇 #DreaminaCPP @dreamina_ai

5. Second Part of the Video Now, since we have a fade to white, we can execute a dissolve effect from white to transition to the second part of the video. In my case, I extracted the transition to white from the first clip and reversed it to recreate the exact same effect and give Seedance the reference. Afterwards, we will use the reference images to create the second part of the video in a single shot. Prompt: REFERENCE MAP: "img" = vid1 (dissolve effect, opening dissolve only) · "27373ff8-c74f-4979-863c-4894866fbae4" = img1 (pale-haired woman in iridescent clothing before a crystalline city) · "c409cd40-4c07-47b3-88dc-650c08456f68" = img2 (ornate silver and gold double-bladed crystal spear, on white) · "783dd18a-c4a2-4311-b5f2-a2b23e679ca8" = img3 (STYLE REFERENCE ONLY, NOT A FRAME: perfectly round water droplets suspended motionless in bright blue air, each throwing sharp starburst flares, dense bokeh field behind) · "b9306d6d-06ab-4b02-b0f9-2c2c3f27ff91" = img4 (STYLE REFERENCE ONLY, NOT A FRAME: faceted ice snowflakes, starburst flares). CRITICAL RULE: img3 and img4 ARE NOT FRAMES OF THIS VIDEO. They are style references only, never reproduced, recreated or shown as a shot. Do not copy their composition, framing or camera angle. Every droplet and snowflake shot is a completely new, originally composed image built only in the spirit of those references. img3 and img4 are never shown full-screen and never appear as an image inside the frame. Cinematic fantasy VFX commercial, Hollywood grade. 15 seconds. Celestial high-key world, strong white backlight, rim light, heavy bloom, prismatic flares, pale blue and silver-white palette turning to brilliant white-gold in the final act. DEPTH OF FIELD IS THE DEFINING LOOK IN EVERY SHOT: extremely shallow focus, wide aperture, only the subject sharp, everything behind it heavy creamy bokeh. 00:00-00:04 - Opens from white and resolves through vid1's dissolve. img1's woman appears in extreme close-up, fully out of focus, a luminous blur. A slow rack focus brings her to perfect sharpness while the camera pushes out and orbits slightly, widening to waist-up. The city behind stays deeply defocused. She appears once. 00:04-00:05:15 - She snaps her fingers and img2's crystal spear materializes floating in the air beside her, four metres long, glowing, arcane glyphs spiralling around it. She reaches out and takes hold of it. 00:05:15-00:07 - Gripping the spear she thrusts it toward the lens, tip aimed straight at camera. Very wide short focal length grossly exaggerating perspective: the tip and her hand loom enormous in the foreground, her face far smaller behind them. The camera dollies fast along the shaft toward her face. Barrel distortion, extreme foreshortening. She casts a spell straight into the lens. 00:07-00:08 - The spell becomes a beam of magical light rushing at the camera: a luminous white-gold shaft built from swirling particles and thin trails of golden smoke, sparks streaming along its length. 00:08-00:10:30 - The beam arrives in a field of perfectly round water droplets in the spirit of img3, hanging completely motionless in the air against luminous blue, each bead crisp and throwing sharp starburst flares, hundreds of them receding into a deep bokeh field. The camera sits inside this suspended cloud and orbits through it at normal speed, weaving between the still drops, parallax sliding them past one another. The magical light threads through the field, its golden particles and smoke curling around each droplet, wrapping them one by one. 00:10:30-00:12 - The light enters the droplets. Each bead ignites from within and turns brilliant gold, no longer water but liquid gold suspended in mid-air, glowing white-hot at the core, then hardens and freezes solid. 00:12-00:13:15 - The frozen golden beads shatter. At the instant of the burst the focus is thrown off completely: the golden snowflake stars in the spirit of img4 are never sharp, appearing only as heavily defocused shapes and glowing bokeh orbs scattering in slow motion across a fully blurred frame. From every burst a thread of white-gold magic escapes and streams outward through the blur. 00:13:15-00:15 - All the threads of golden magic converge and weave together, and the title is revealed out of that light against the already defocused field: bold sans-serif lettering reading exactly "SEEDANCE 2.5", spelled S-E-E-D-A-N-C-E, forming from brilliant white-gold energy with bloom and lens flares. The lettering is the only sharp element in the frame, razor crisp against a completely melted background of golden bokeh, held to the very last frame. Style: fantasy, celestial, photorealistic VFX, god rays, backlit, lens flares, bokeh, slow motion, golden particle magic. Camera fluid, eased. 4K.

4. Generate the Beginning of the Video: 3D Image Panel to Slider Carousel We will use the video generated in step three and attempt, with the image from the first frame, to generate the beginning of the video. Prompt: TITLE CARD, THE FIRST THING THIS VIDEO SHOWS: bold sans-serif lettering fully present and legible in the very first frame at 00:00, floating in the foreground, sharp and in focus from the start. First line reads exactly "SEEDANCE 2.5", spelled S-E-E-D-A-N-C-E. Second smaller line beneath reads exactly "Up to 50 references". Dark charcoal type with the words "50 references" in bright cyan blue. Crisp, sharp-edged, perfectly legible. It sits weightlessly over the moving background and fades out cleanly just before 00:04. High-budget SaaS motion design commercial. Total duration 14 seconds. FIXED TIMING, NON-NEGOTIABLE: reference vid1 starts at exactly 00:04 and plays for ten full seconds, ending at 00:14. Vid1 occupies the majority of this video. It runs from its very first frame to its very last frame as one single continuous playback, in its original order and at its original pace, with all ten seconds present. Its closing seconds play out completely and at full length. The generated portion is only the first four seconds and it must finish by 00:04 so that vid1 begins on time and has room to run whole. THE ENTRY PASSAGE (00:00–00:04), starting from reference img1. It moves briskly and covers its ground within four seconds. Endless bright white space, pale blue, periwinkle and silver-white palette, soft fog, glossy floor with gentle reflections. RACK FOCUS: the passage opens completely out of focus, a soft creamy blur of luminous shapes and glowing colour with no legible detail. Over the first three seconds the focus pulls in gradually and continuously, edges resolving and card borders sharpening until by 00:03 the image is perfectly sharp and holds that way. A smooth cinematic lens rack, eased at both ends. The title lettering stays sharp throughout, unaffected by the blur behind it. It opens on a tight crop of ONE continuous inclined 3D plane built from luminous image cards: only four large cards are visible, and the same grid carries on beyond all four edges of the frame, with further rows above and below and further columns left and right. The plane eases into a slow rotation, tilting toward a more vertical angle while staying inclined, floating weightlessly. A dolly-out then eases in and pulls back steadily from this identical surface. More of the same grid enters the frame in every direction while the four original cards stay visible and shrink, joined by hundreds of smaller cards from the same wall. It remains one single continuous plane at all times, the very same surface throughout, simply seen wider, until the full immense mosaic fills the frame edge to edge, intact and complete, light sweeping across it. A wave of dissolution then sweeps inward from the top edge and the bottom edge simultaneously, cards fading softly into the white void band by band, the wall thinning until a single central horizontal strip remains. That strip enlarges into a BROAD CAROUSEL BELT of large 16:9 panels standing ONE PANEL TALL, a single chain joined only left to right, with open white void above it and open white void and reflective floor below it. Every panel shares the same top edge and the same bottom edge. The belt is slightly tilted in 3D, its ends converging toward vanishing points. The belt travels from left to right, accelerating continuously into a high-speed stream, the camera panning right with steadily increasing angular speed, still gaining momentum as the passage ends. THE HANDOVER INTO VID1: the entry passage ends at 00:04 on a frame matching vid1's first frame exactly in belt speed, camera pan velocity, framing, angle, direction, light and focus, fully sharp on both sides of the join. The movement carries straight through as one continuous gesture and vid1 takes over seamlessly, then runs all the way to its end. The video finishes on vid1's own final frame, the white-out that closes it, arriving only after all ten of its seconds have already played through. The last moment is pure blinding white filling the entire frame, fully blown out edge to edge, held to the very last frame. Panel effects in the entry passage: broad specular reflections rake diagonally across the surfaces like light sliding over polished glass, flaring brighter as they travel and momentarily blowing whole panels out to pure white before easing back. Panel borders glow with a fine luminous rim and prismatic fringing in peach, icy cyan, lavender and pale gold. Bloom is strong yet smooth, holding steady. Colours stay within the scene's own cool spectrum. The entry passage uses strict ease-in and ease-out curves, with the belt phase as one unbroken accelerating ramp. All motion stays fluid and precisely controlled, motion-design grade. Photorealistic 3D render, minimal, elegant, 4K.

3. Generate a Dynamic Image Carousel and a 3D Box with Dissolve We will divide the video into 2 parts. In the first part, we show the interface ending in a fade to white. We will also divide this first part into 2. We'll start by making the second one, which is an image carousel that ends on the interface and finishes with a fade to white. Video prompt: High-budget SaaS motion design commercial, continuation shot. Reference vid2 provides the environment: an endless white void with tall translucent glass slabs, a vertical shaft of backlight, volumetric haze, polished reflective floor, high-key pale blue, periwinkle and silver-white palette, with soft slow-morphing light leaks in peach, amber, icy cyan and lavender drifting across the frame, heavily feathered, out of phase with one another. Reference img1 provides the interface design, its layout, proportions and its two image carousels. PHASE 1 — The shot opens already in motion: a wide horizontal RIBBON of large 16:9 image panels, exactly one panel tall, races from left to right at high speed across the vid2 environment, the camera panning right to follow it. The ribbon and the camera both decelerate together with a long, smooth motion-design ease-out as they arrive at a new area of the white void, where the translucent interface panel from img1 materializes floating in space. The travelling ribbon merges seamlessly into the interface, becoming the upper-left carousel seen in img1: it enters from the right side of the frame, sweeps leftward across the composition and eases out to settle along the left edge, still flowing. Moments later a second identical ribbon forms in the lower part of the frame, emerging from the centre-bottom of the interface and extending toward the right. Both carousels run in the same direction, images travelling left to right at speed. The upper-left ribbon terminates at the centre of the composition with a soft dissolve, its panels fading out where they meet the interface; the lower-right ribbon begins at the centre-bottom with the same soft dissolve, its panels fading in from nothing. The camera glides smoothly through space and settles into its final framing, with the translucent glass interface panel floating centred, its empty central image slot visible as a clear frosted glass rectangle with glowing edges. PHASE 2 — The camera is now nearly static. Both carousels continue flowing while, one after another, four image thumbnails appear inside the empty central slot of the interface, materializing sequentially from the centre outward with a soft opacity fade-in, each one a beat apart. As the images populate, both carousels gradually and gently decelerate, their speed easing down along a smooth motion-design deceleration curve until they are barely drifting. Throughout, the interface panel floats weightlessly in space, its inclination shifting by an extremely subtle amount, an almost imperceptible tilt, as if suspended in air. PHASE 3 — A tunnel of light builds from behind the interface: a burned-out light leak expands rapidly toward the camera, bloom flooding outward, the glass and the panels dissolving into pure luminance until the entire frame turns completely white. Clean white-out transition. Style of carousels and interface borders: dynamic glow travelling continuously along every edge, tracing the borders of the panel and of each image card like a running highlight, light glowing through translucent frosted glass. Broad specular reflections rake diagonally across the surfaces, momentarily blowing sections out to pure white before easing back. Prismatic fringing in peach, icy cyan, lavender and pale gold. Bloom is strong yet smooth. No magenta, no saturated colors. All motion follows strict ease-in and ease-out curves: every movement starts slowly, accelerates to gather speed, then decelerates softly. Motion-design grade smoothness, fluid and precisely controlled, no shake, no cuts, no handheld feel. Photorealistic 3D render, minimal, elegant, 4K.

2. Generate the Video Images Give me an image of a woman, sci-fi fantasy style, that matches the color palette of the reference image, rim light, medium shot from the waist up, pale face, and long white hair. Prompt: Create an image of a white background with a double-ended spear where each end is the tip of a magic scepter (identical) and the weapon has many baroque-style ornamental details and then ends in magic gold circles with golden yellow light like an orb. Give me 3 versions. The style will be based on the reference image. Prompt: An image in the same style but of frozen water flakes (in the style of their microstructure) in this style, floating, celestial background. Prompt: An image of water drops, detail shot, extreme detail, where you can appreciate the light and sparkles and lens flare passing through it.

1. Generate the Interface Style Frames Generate the central images that will act as anchor points for the key moments of the video. In my case, I took a screenshot of the official Dreamina website interface. I used GPT-2 for all the images in this video. First, the interface frame, prompt: The uploaded image is from the Dreamina website. The logo is blue, give me a color palette for a high-budget motion design video for this brand. It would be a background with light leaks and the image reference box with a slider like a tilted screen in 3D space. And then with the interface I got an image, prompt: Give me, based on this style, an entire 3D image panel, with 16:9 images, flat tilted style formed by image panels and a close camera. Finally, I removed all the elements to generate the background, prompt: Give me the image of just the background without the interfaces (3 in total) all 16:9.

Seedance 2.5 - Advanced Workflows Series 2. PREMIUM SAAS MOTION DESIGN COMMERCIAL Seedance 2.5 is designed to get the maximum control out of it using references. The prompt adherence is extremely good, and its "section-based" design style allows you to give specific instructions independent of camera, style, effects, and action. Blur and light effects, animations with ease in/ease out timing—this model can handle it all! I designed this entire commercial in about 5 hours without touching After Effects for a single moment. Only art direction and artificial intelligence. Workflow + Prompts 👇

SEEDANCE 2.5 IS HERE! Seedance 2.5 - Advanced Workflows Series 1. 30 Image References Seedance 2.5 is here — now supporting up to 30 image references. Animation quality has taken a clear step forward, with smoother motion and a much stronger ability to replicate Japanese 2D animation styles and their signature movement. Prompts 👇

Disney fired all its 2D animators, illustrators, and closed its 2D animation department in 2013. The company that invented pencil magic killed its own legacy. The paradigm shift actually started in 1995 with the massive success of Pixar's 'Toy Story', the first-ever 3D animated feature. By the early 2000s, audiences were flocking to CGI hits like 'Finding Nemo' and DreamWorks' 'Shrek'. 3D became the shiny new standard, while Disney's traditional films like 'Treasure Planet' (2002) were massive box office failures that lost the studio tens of millions. 👇(1/4)

Disney fired all its 2D animators, illustrators, and closed its 2D animation department in 2013. The company that invented pencil magic killed its own legacy. The paradigm shift actually started in 1995 with the massive success of Pixar's 'Toy Story', the first-ever 3D animated feature. By the early 2000s, audiences were flocking to CGI hits like 'Finding Nemo' and DreamWorks' 'Shrek'. 3D became the shiny new standard, while Disney's traditional films like 'Treasure Planet' (2002) were massive box office failures that lost the studio tens of millions. 👇(1/4)

4. Generate the Animations Separately in Vidu Q3 Vidu Q3 is a video model that has worse prompt adherence than Seedance 2.0 but, in exchange, it generates much more expressive animations. Despite having a fairly high percentage of generations with artifacts, around 10% of the generations contain extremely high-quality and highly expressive animations that Seedance 2.0 can rarely reproduce. To obtain the highest-quality AI-generated cut, we will combine the power of both models, allowing Vidu to generate highly expressive animations of characters alone or separate elements while trying to maintain a consistent shot, and using Seedance 2.0 for the less expressive animations, where we can take advantage of its prompt-adherence capabilities. In rubber hose style, squash and stretch, and highly expressive anime genres such as battle shōnen, Vidu Q3 animations are far superior to those of Seedance 2.0. Vidu Q3: --- Control ++++ Expressiveness Crocodile and Cat Prompt: SINGLE SHOT, no multicut. NO CHANGE OF CAMERA ANGLE Old-school cartoon western, rubber hose animation, strong squash and stretch, bouncy cartoon physics, elastic legs, looping hat bounce, expressive facial animation, exaggerated eyebrow motion, animated mouth acting, stylized gunfire, hand-drawn 2D animation, slight ground smear. Horses Prompt: SINGLE SHOT, no multicut. NO CHANGE OF CAMERA ANGLE Old-school cartoon western, rubber hose animation, strong squash and stretch, bouncy cartoon physics, elastic horse legs, riders bouncing out of sync, looping hat bounce, expressive facial animation, exaggerated eyebrow motion, animated mouth acting, stylized gunfire, hand-drawn 2D animation, slight ground smear. Main Character Prompt: SINGLE SHOT, no multicut. Use the provided image as the first frame. The camera tracks smoothly with the square cowboy, preserving the same composition, dutch angle, character size, and framing throughout while he runs continuously preserving the framing an direction of the first frame with extreme LIGHT squash and stretch: his square body compresses with every landing, stretches tall during each bounce and ripples elastically as it recovers, while his arms and legs bend and extend into exaggerated rubbery shapes while his cowboy hat repeatedly rises above his head and falls back onto it in a continuous desynchronized loop. His frightened face stays highly animated while looking behind him: His pupils make small nervous movements, his eyelids twitch subtly, and his mouth trembles and changes shape slightly while he runs. The ground scrolls beneath him with slight horizontal smearing, and each footstep raises a small puff of dust. He is alone, and nothing pursuing him is visible. The ground scrolls beneath him with slight horizontal smearing, and each footstep raises a small puff of dust. He is alone, and nothing pursuing him is visible. Old-school cartoon western, rubber hose animation, extreme full-body squash and stretch, elastic twisting body deformation, highly expressive facial animation, exaggerated mouth and eye movement, unrealistic cartoon physics, hand-drawn 2D animation, slight ground smear.

3. Separate the Layout into Layers with GPT 2 to Create Different CELs In the same ChatGPT chat, you can use a prompt that generates several images by separating each of the characters together with the background. These images will work as independent cels—layers with separate animations—which you will later recompose into the final shot. GPT 2 Prompt: Make 3 images, keeping the background but: 1. only the cat on the right and the crocodile on the left 2. only the horses in the background with the cowboys on top 3. only the main character in front keep the original size of the characters and the background and the framing and the position, only remove the elements that are left over to form the image keep the ar 16:9

2. Generate the Final Layout in Nano Banana / GPT 2 With the different characters and an initial frame that clearly captures the camera angle and the relative positions of the main character and the background, you can create a final edit that will serve as your first frame. Nano Banana Prompt: cartoon, western, quircky, ruber hose, a monster in a convoy, use @[Image 4](image_4) as the main reference for the running monster: a quirky little western monster with a square head and simple body, running in the foreground, dutch angle, eyes wide open, mouth open with teeth showing, comic panicked expression, short skinny limbs, cowboy hat bouncing as he runs, same angle and same character energy as [img4], but fully in color. In the background, a group of chasing characters follows behind him. Use @[Image 1](image_1) as reference for the 2 mustached cowboys riding 2 frightened horses: the cowboys have big mustaches and exaggerated old-school western expressions, and the horses have big scared eyes and alarmed expressions. Place them in the background behind the monster. Use @[Image 2](image_2) as reference for a funny cowboy crocodile carrying a shotgun, chasing behind the monster. Use @[Image 3](image_3) as reference for a cowboy cat with a pistol, also chasing behind the monster. The background should follow the visual style of @[Image 2](image_2), with a simple old-school cartoon desert setting, cacti, open western landscape, and clean stylized shapes. Classic old-school cartoon western, exaggerated motion, lively, playful, vintage theatrical animation, expressive poses, whimsical and energetic

1. Generate the Characters and Concept Art in Midjourney The best workflow for generating high-quality images is to use Midjourney for the style and then Nano Banana / GPT 2 to create the final composition and clean up inconsistencies. I wanted an image with a dutch angle in a rubber hose-style animation of a little monster being chased by a crocodile, two cowboys riding horses, and a cat. To do this, I generated a base image for the main character and then another three for the secondary characters: Midjourney Prompt — Main Character: cartoon, western, quircky, ruber hose, a monster in a convoy, quirky little western monster with a square head and simple body, running in the foreground, close-up, dutch angle, eyes wide open, mouth open with teeth showing, comic panicked expression, short skinny limbs, cowboy hat bouncing as he runs. In the background, 2 cartoon cowboys on horses with giant eyes chase after him with pistols, alongside a goofy crocodile with a shotgun and a Native American archer firing arrows. Classic old-school cartoon western, exaggerated motion, lively, playful, vintage theatrical animation Midjourney Prompt — Horses: cartoon, western, quircky, ruber hose, running toward the viewer in a dramatic dutch angle shot, expressive action pose, 2 mustached cowboys riding only 2 frightened horses on the left side of the image in the background, far from the camera, both clearly mounted on horseback, expressive old-school western characters, the horses look scared with wide eyes and alarmed expressions, horses with big eyes, Rubber hose cartoon style, looney old western energy, dynamic chase composition, dusty frontier road, whimsical and exaggerated Midjourney Prompt — Cat: cartoon, western, quircky, a cowboy cat with a hat tin the desert, small body big head, stylized, sharp eyes and head, angry face, full body view, tex avery style Midjourney Prompt — Crocodile: cartoon, western, quircky, a crocidile like monster with a cowboy hat tin the desert, tex avery style

Seedance 2.0 - Advanced Workflows Series 14. Compositing and Layering Some shots may have many characters that need specific movements. When we start adding characters, it becomes very difficult to achieve the exact animation, gesture, or movement we are looking of all of them in a single generation. In addition, the more characters are added, the more rigid Seedance 2.0 animations become in order to maintain consistency. This is especially visible in styles with squash and stretch and rubber hose animations, where the characters deform their body parts as if they were made of rubber, or in highly expressive movement animations in anime. For these cases, it is better to work on each animation separately and recompose it at the end into a video. Workflow + Prompts 👇

3. Separate the Layout into Layers with GPT 2 to Create Different CELs In the same ChatGPT chat, you can use a prompt that generates several images by separating each of the characters together with the background. These images will work as independent cels—layers with separate animations—which you will later recompose into the final shot. GPT 2 Prompt: Make 4 images, keeping the background but: 1. only the cat on the right and the crocodile on the left 2. only the horses in the background with the cowboys on top 3. only the main character in front keep the original size of the characters and the background and the framing and the position, only remove the elements that are left over to form the image keep the ar 16:9

2. Generate the Final Layout in Nano Banana / GPT 2 With the different characters and an initial frame that clearly captures the camera angle and the relative positions of the main character and the background, you can create a final edit that will serve as your first frame. Nano Banana Prompt: cartoon, western, quircky, ruber hose, a monster in a convoy, use @[Image 4](image_4) as the main reference for the running monster: a quirky little western monster with a square head and simple body, running in the foreground, dutch angle, eyes wide open, mouth open with teeth showing, comic panicked expression, short skinny limbs, cowboy hat bouncing as he runs, same angle and same character energy as [img4], but fully in color. In the background, a group of chasing characters follows behind him. Use @[Image 1](image_1) as reference for the 2 mustached cowboys riding 2 frightened horses: the cowboys have big mustaches and exaggerated old-school western expressions, and the horses have big scared eyes and alarmed expressions. Place them in the background behind the monster. Use @[Image 2](image_2) as reference for a funny cowboy crocodile carrying a shotgun, chasing behind the monster. Use @[Image 3](image_3) as reference for a cowboy cat with a pistol, also chasing behind the monster. The background should follow the visual style of @[Image 2](image_2), with a simple old-school cartoon desert setting, cacti, open western landscape, and clean stylized shapes. Classic old-school cartoon western, exaggerated motion, lively, playful, vintage theatrical animation, expressive poses, whimsical and energetic

7. Render the video in Seedance 2.0 Use the restylized image of the first frame and the video with the mask so that Seedance 2.0 generates the complete clip. After playing with chatgpt and improving the prompt this is the one I used: SINGLE SHOT, no multicut. Use the video as a reference for the camera work, the push-in and the general blocking of the elements in the scene. The image1 corresponds to the first frame and must serve as a visual guide for the initial composition of the shot. The red objects that rotate in the video represent asteroids. Use their movement, spin and flotation as a reference to animate the asteroids in space. The green square represents only the approximate position, scale in shot and spatial location of the astronaut regarding the camera. Use it only as a placement marker inside the framing. Completely replace the green square with an astronaut of image1 with helmet and spacesuit. The square must not appear in the final result. It must serve solely as a reference of where to place the astronaut in the scene. The scene shows asteroids floating in space while the camera does a push-in towards an astronaut with helmet and spacesuit. The astronaut floats in zero gravity and occupies the position marked by the green square, with a very subtle, slow and controlled body animation. The astronaut does not perform effusive or agitated movements. Remains floating with very slow movements, almost suspended, maintaining a sensation of passive drift in space. The arms move only very slightly, with soft gestures as if they were displacing backwards by inertia or trying to stabilize with minimal effort. The body remains calm, with slow and natural micro-adjustments of balance in zero gravity. The legs remain completely still during the whole video. There is no leg movement whatsoever. All the body animation must concentrate solely on small, slow and subtle movements of the arms. The head also moves with a lot of containment. If there is nervousness, it must be felt only in a subtle way, through small slow gestures, not through abrupt movements. The general sensation must be of an astronaut lost and floating adrift, but expressed with a minimal, slow and credible performance. Very important: maintain the same relative position in frame, scale in shot and relationship with the camera indicated by the green square, but substitute that marker with the complete astronaut. The green square is only a location guide and must disappear completely in the final render. Important: backlit illumination, not frontlit. An intense light from the background illuminates the scene, creating silhouettes, strong rim light on the asteroids and the astronaut, dramatic spatial glow and cinematic contrast. Photorealistic style, cinematic, with a background of stars, spatial depth and credible flotation movement.

6. Restylize the first frame of the viewport: Once you have your midjourney image, you will use it to restylize the first image of the generated viewport. This way you will generate a rendered image with the materials and the light of your own video, which will serve so that Seedance 2.0 renders the scene following this composition of light and elements. Prompt: Use Image1 as a spatial and blocking reference and create an image replacing the objects red with astreroids, with a space background and a strong backlit using Image2 as a reference.

5. Generate an image in Midjourney as a style reference: Take a screengrab of your viewport (with dummy, without mask). Give it to chatgpt and tell it that based on the geometry of the image it generates a prompt of the scene in midjourney. In this case, I asked it to generate a photograph of an astronaut in space floating between asteroids, with a color palette and lighting that I liked: Photorealistic dark outer space scene with numerous floating asteroids drifting through a deep black cosmic environment, large foreground asteroids and smaller distant rocks, highly detailed rocky surfaces with craters, dust and mineral texture, strong backlit illumination only, no front lighting, intense luminous glow from behind the asteroids, dramatic rim light outlining the rocks, striking spatial light effects, radiant beams, subtle glowing particles, deep shadows, high contrast, a wide star-filled background, and a distant astronaut wearing a helmet, very small in scale, floating quietly among the rocks, strong sense of emptiness, depth and silence, realistic space photography, ultra detailed, atmospheric, high dynamic range

4. Generate a character mask in After Effects Seedance has a problem: when you give it precise information it tends to adhere to it, and limits its imagination capabilities. Seedance is a tool designed to have high adherence and cohesion. This causes that, when we give it a "stiff" dummy in a scene and we ask it to replace it, Seedance will imitate the original position of the dummy in the video. This occurs also with images: if you provide completely fixed frames, seedance will have difficulties in imagining how to connect the intermediate steps from one image to another. What to do? OMIT INFORMATION. We will generate a solid with a green square in After Effects that makes a mask of the character. This way we will maintain the proportions and position of our character, but we will avoid giving information about the bones, the body structure and the pose, thus we will allow seedance to use all its power to imagine the animation of the character.

3. Generate a cinematic Once the level is generated, you will be able to play inside it as if it were a game. For our cinematic, we only need the astronaut to move in one direction in the environment. Upon hitting play, we will be able to handle our dummy and we will only have to move it in one direction. We will record this gameplay. Afterwards, we will place a virtual camera that focuses directly on the astronaut and we will animate its movement inside this recorded "take" of the gameplay.

2. Run Fable 5 to build a game level Once the MCP is configured, you only have to give instructions inside Unreal Engine so that Claude takes control and generates what you want. Prompt: Generate a zero gravity level, without a floor. The background will be a flat blue color. Create 20 asteroids formed by basic shapes, with red color, that float in space. They will make rotation movements on themselves very slowly and will also move slightly in a random direction in space. Create also a dummy with basic shapes of green color, playable, that also has zero gravity. This dummy will also float and will not brake automatically when handled, in such a way that it maintains an inertial movement when it moves in one direction. The playable character will be located right in front of the asteroid cluster.

Seedance 2.0 - Advanced Workflows Series 12. Cinematics with Unreal Engine and Claude Fable 5 Fable 5 is an absolute beast model that can take control of Unreal Engine and create videogames. We can build scenes as a little video game with physics and assets and configure a playable character to create a cinematic scene. In this example I created a scene of an astronaut lost in space floating between asteroids. I created a zero gravity environment and made assets like cubes and spheres representing asteroids interact in this environment. I recorded the gameplay and later place a camera to build the cinematic. Workflow + Prompts 👇

4. Remove the Chroma Key in After Effects After Effects has a very good default chroma keyer called Keylight 1.2: Select the animation layer -> Effect -> Keying -> Keylight 1.2 -> Select the brush and click on the green background of the image.

3. Generate the Animations in Seedance I generated the mini animations in Seedance one by one, with simple prompts describing the action. Prompt: STATIC CAMERA A skeleton puts viruses inside a net, and then takes them to a bag and stores them. DO NOT change the framing, keep the camera static, do not zoom in or zoom out. Afterward, I generated the animation for the vignette. Prompt: STATIC CAMERA A pterodactyl and an office worker laugh out loud. Suddenly the skeleton from the image enters while both look at him suspiciously. The skeleton stands next to them and then the three of them continue laughing again. DO NOT change the framing, keep the camera static, do not zoom in or zoom out.

2. Generate Images with Green Background I asked ChatGPT to give me suggestions for everyday, funny actions that had to do with the theme of the ad (cybersecurity). I suggested it give me the character pushing alert buttons, eating viruses, or looking at a letter screen, and then doing normal tasks like running, or playing with cubes. Once I had all my actions, I generated the images in Nano Banana Pro with simple prompts, using a green background: Prompt: Generate an image of the character in a green screen chroma key backgound #00FF00 doing this action "specific action"

1. Generate a Layout I inputted the image of My Neighbor Totoro into ChatGPT and asked it to make an image in the same style, but with the brand's characters. Prompt: Generate an image with a similar layout as image1, use the pterodactil of image 2 and the oficinist of image 3 and set them in a bacground in the same style as image 1, laughing one next each other, at the left of the image. The background will be a soft modern purple gradient In the horizontal rows at the top and bottom, put the charactesr from iamge 4 and image 5 in alternate positions, next to each other. doing comic poses and actions (pushing buttons, eating, cathing viruses, etc) Give me 4 differente images AR 16:9 Afterward, I asked ChatGPT to keep only the background and the vignette so I could animate them separately.

Seedance 2.0 - Advanced Workflows Series 12. Integrate AI Animations with Green Screen in After Effects A very powerful tool for integrating different animations is to generate them with a fixed camera and a green background. In my last project, an advertising campaign for Torq, I was asked to create a credits screen referencing "My Neighbor Totoro." In the original film, the images are not animated, but it occurred to me that it could be very dynamic to integrate mini animations in two sliders moving in opposite directions. This is the magic of Seedance 2.0: you can generate different animations by layers and integrate them later in After Effects to create mind-blowing effects. Actually, this is how it's done in traditional animation; different layers with different animations are integrated into the final image. Workflow + Prompts 👇

This project is a true hybrid production: we merged AI animation with After Effects motion design, digital illustration, and hand-drawn cel animation. It's a campaign for Torq, a cybersecurity company, produced by Thinkmojo. A wonderful highlight for me was seeing cel-animated the opening shot, based on an AI layout I designed for the spot! A huge thank you to the teams at Torq and Thinkmojo—it was an absolute pleasure working together!

Cartoon Hero 3.0 is launching! Registrations are open for this week only! (launch date 2nd to 8th June) This is where I started learning AI animation. It’s not just a course; it’s a community of over a thousand AI animators featuring: -Dozens of courses -Exclusive member contests with thousands of dollars in prizes -Discord server with professionals from the film, advertising, and animation industries -Job opportunities I am an active member of this community. It has not only taught me a lot, but it has also opened doors for professional work. Don't miss out on this opportunity—join our community today 👇

This is an extremely good genius short film from Javier Ara, an spanish artist, called Dreaming a Whole Life (2011) You can watch it on youtube: piped.video/S-G2lG-o_uk?is=fhP3…

I'm finishing up Jujutsu Kaisen and it's a masterpiece of such caliber (layouts, animation, lighting, VFX) that it made me want to make some battle shonen scenes. I started this scene, but I suddenly got hit with a ton of work and couldn't finish it. So, I'm going to post what I have so far—if you want me to finish the fight, let me know in the comments or repost! Since my Japanese friends were the ones who made my last post take off, I've decided to post this during their timezone to show my appreciation. (ありがとうございます) If I get 100 reposts, I promise to make a part two and put real effort into a great fight scene. #DreaminaCPP @dreamina_ai

4. Select the Best Shots From Multiple Generations of Video Analyze the multiple generations and extract the most beautiful shots you can find. In my case, I kept 16 shots, which I later used to make a small storyboard. I had a total of 16 shots, which I decided to animate in 2 different generations: 8 + 8. In each generation, I provided the list of shots, and in each shot's section, I referenced the specific image, so that Seedance would replicate the exact same shot I had obtained in the previous videos.

3. Get Shots from Seedance Use your new agent to get beautiful shots. Input a source image and ask the agent to provide you detailed text prompts with beautiful compositions of the scene. I asked for 20 variations. I used the exact output of the agent and fed it into Seedance. I only provided a beautiful Source Image. Prompt: 10 shots. Duration per shot: 0.4s. Transition: hard cuts between all shots. Shot 1: cinematic anime layout, lonely blonde anime girl standing still in a massive crowded Tokyo crossing at sunset, extreme wide shot, high camera angle, horizon very high, huge wet reflective ground, overwhelming urban architecture, giant LED screens towering above, crowd flowing left to right while girl faces opposite direction, deep atmospheric perspective, foreground silhouettes heavily blurred, cinematic bokeh, volumetric golden light, emotional isolation, two point perspective with converging lines toward the girl, realistic anime movie composition, subtle melancholy, Makoto Shinkai inspired cinematic framing Shot 2: anime cinematic voyeur composition, lonely girl seen through a dense crowd, shot through human silhouettes, telephoto lens compression, tiny visual opening revealing only her face and red scarf, heavy foreground blur occupying most of frame, cinematic depth layering, realistic urban anime atmosphere, warm golden city lights, surveillance feeling, emotional tension, shallow depth of field, subtle eye contact, mature anime movie framing Shot 3: cinematic anime loneliness composition, blonde anime girl tiny in lower left corner, enormous negative space dominating frame, abstract glowing city bokeh, low camera angle, horizon below frame, dreamy urban haze, emotional emptiness, warm sunset bloom, minimal composition, poetic anime film framing, isolated human presence in overwhelming atmosphere Shot 4: perfect one point perspective anime street composition, lone girl centered in crowded neon city, vanishing point directly behind her head, ceremonial cinematic symmetry, crowded urban corridor, warm atmospheric glow, balanced architecture, emotional stillness inside chaos, subtle asymmetrical pedestrians breaking perfect geometry, anime movie visual language Shot 5: dramatic low angle anime city composition, tiny lonely girl surrounded by gigantic skyscrapers, three point perspective, strong vertical convergence, horizon below frame, overwhelming urban scale, cinematic neon sunset, oppressive architecture, emotional vulnerability, anime cinematic realism, atmospheric crowd silhouettes Shot 6: anime cinematic reflection shot, lonely girl reflected on a rainy shop window at sunset, layered composition with city reflections and interior lights, emotional introspection, soft diagonal framing, realistic anime movie atmosphere, urban melancholy, warm and cold light contrast, cinematic depth and reflections Shot 7: extreme close-up anime eye shot, emotional blonde girl in crowded city, abstract urban bokeh background, hair crossing frame diagonally, cinematic shallow depth of field, melancholic atmosphere, emotional realism, warm sunset lighting, highly expressive anime film composition Shot 8: cinematic anime emotional separation, two characters divided by a moving crowd in a Tokyo street at sunset, strong visual barrier made of human silhouettes, emotional distance, layered depth composition, warm cinematic glow, anime movie realism, melancholy urban atmosphere Shot 9: documentary style anime street shot, imperfect framing, candid cinematic composition, lonely girl partially cut by frame edge, natural crowd movement, handheld feeling, subtle dutch angle, realistic urban atmosphere, emotional realism, shallow depth of field, cinematic anime photography Shot 10: psychological horror anime composition, lonely girl in crowded city with ominous empty space behind her, threatening unseen presence, dark foreground silhouettes, cinematic low angle, tension outside frame, warm urban lights contrasting fear, atmospheric anime thriller framing 10 shots. Duration per shot: 0.4s. Transition: hard cuts between all shots. Shot 11: epic cinematic anime city shot, low angle ceremonial composition, lonely girl beneath monumental glowing urban architecture, divine sunset backlight, symmetrical skyscrapers, mythic urban atmosphere, emotional awe, anime movie visual grandeur Shot 12: iconic anime silhouette composition, lonely girl in full backlight inside crowded city, pure readable silhouette, glowing scarf and hair edges, abstract urban crowd, cinematic golden haze, emotional minimalist anime framing Shot 13: telephoto compressed anime city shot, overwhelming dense crowd surrounding lonely girl, no breathing space, emotional suffocation, cinematic long lens compression, warm urban bokeh, realistic anime atmosphere, psychological isolation Shot 14: dynamic anime crowd composition, everyone moving rapidly in one direction while lonely blonde girl walks slowly against the flow, strong directional contrast, diagonal movement lines, cinematic urban sunset, emotional resistance, layered crowd depth, anime movie realism Shot 15: claustrophobic anime crowd shot, giant blurry human silhouettes invading foreground, lonely girl barely visible between bodies, oppressive social pressure, cinematic urban atmosphere, emotional suffocation, layered depth composition, realistic anime movie framing Shot 16: poetic anime sunset composition, lonely girl standing in city with enormous glowing sky dominating frame, very low horizon line, lyrical cinematic atmosphere, emotional contemplation, soft clouds and urban silhouettes, warm cinematic anime lighting Shot 17: psychological anime city shot, subtle dutch angle, vanishing point near frame edge, emotional instability, lonely blonde girl in crowded urban environment, distorted perspective tension, cinematic sunset glow, expressive anime movie composition Shot 18: top down anime crowd composition, lonely girl identifiable only by red scarf among massive urban crowd, abstract human patterns, cinematic overhead shot, emotional anonymity, warm city glow, realistic anime movie atmosphere Shot 19: cinematic anime side profile shot, lonely blonde girl staring into huge negative space outside frame, emotional anticipation, off-screen tension, atmospheric urban sunset, soft crowd silhouettes, poetic anime film composition Shot 20: masterpiece cinematic anime composition, lonely blonde girl standing still in crowded Tokyo street at golden hour, medium wide shot, human eye level camera, moderate telephoto lens, soft natural perspective with off-screen vanishing points, strong foreground silhouettes blurred into abstract shapes, flowing crowd moving diagonally while girl remains still, glowing red scarf as secondary focal point, warm volumetric sunset light, emotional urban melancholy, layered cinematic depth, realistic anime movie atmosphere, subtle wind in hair, poetic visual storytelling, Makoto Shinkai and Satoshi Kon inspired framing

1. Get Cinematic Compositions You might be an expert storyboarder, a brilliant layout artist, or a very good photographer. In that case, you just need to use your knowledge to craft detailed descriptions of beautiful shots. But if that's not the case: DON'T WORRY! ChatGPT, Gemini, and Claude have options to build small agents specialized in specific tasks. In my case, I tried to create a LAYOUT or STORYBOARD artist within my AI. To do this, I first asked ChatGPT, using its 5.5 Thinking model, to make an extensive report on composition rules, perspective, etc., that a professional storyboard artist knows perfectly. ChatGPT gave me back a report that TAKES UP 12 PDF PAGES!! (I uploaded it below) Afterward, I created a custom GPT (a small assistant) that I named ANIME LAYOUT MASTER. I established rules for its behavior and attached the 12-page PDF with the knowledge it needs to help me.

Seedance 2.0 - Advanced Workflows Series 11. Cinematic Camera Angles through Video Gen + Specialized Storyboard AI Agent Are you already getting camera shots and angles through Nano Banana Pro? Take the next step. Seedance 2.0 has a better spatial understanding of the scene than image generation models. Because it is built to produce cinematic clips, it achieves more beautiful shots and angles than image models like Nano Banana, which are primarily made for image editing. As an added bonus, Seedance renders the entire space of the scene, so you can get tens of shots with complete spatial and element consistency in a single generation. Create a specific AI Agent to help you with storyboarding, and feed the input into Seedance to get the most cinematic shots. You can access Seedance 2.0 now on insMind (link at the end of the thread). Workflow + Prompts👇

2. Use GPT Image 2 to re-render the image With the provided prompt, upload the image and ask it to restore, enhance, or upscale it. You can also use a simpler prompt if you only want to upscale and there are 2 or 3 specific defects you have pinpointed and want to fix. When you only want to improve the resolution, you can add character references so it doesn't modify the textures. Add phrases and images that help preserve the style. Prompt (upscale with character references): Upscale this image. Preserve colors and lightning and color grade. Use images 2, 3 and 4 as a reference to render look and textures of the characters. Prompt (image restoration of character sheet): Restore and enhance this character reference sheet while preserving the exact original composition, layout, character design, proportions, colors, lighting style, shading, materials, and artistic style. Keep all views and portrait close-ups exactly in the same positions and preserve the original pastel pink-to-blue hair gradient, skin tones, clothing colors, boot glow, and neutral gray background tone. Clean up all visible compression and quality defects, including banding, posterization, debanding artifacts, mild blur, edge halos, ringing, dirty edges, compression smearing, and loss of fine detail. Smooth the tonal transitions in the background, hair, skin, and clothing so gradients become clean and continuous, but do not change the original colors or lighting. Restore fine detail in the hair strands, bangs, clothing fabric, denim shorts, fishnet gloves, straps, buckles, chain, stitching, and boot laces, making them cleaner and sharper without changing the design. Improve edge quality around the hair silhouette, clothes, boots, accessories, and close-up portraits, removing artifacts while keeping the original shape exactly. Keep the image looking like the same polished 3D stylized character sheet, only cleaner, higher resolution, and artifact-free. Do not redesign any element, do not alter the pose, expression, anatomy, proportions, texture style, or material response. The result should be a high-quality restoration/upscale of the same image, with preserved original colors, preserved proportions, preserved lighting mood, and preserved texture style. Prompt (realistic image): Upscale this image. Preserve colors and lightning and color grade. Use images 2, 3 and 4 as a reference to render look and textures of the characters. Prompt (anime): Upscale these images. Preserve the anime, cel-shaded style. Do not add realistic textures. High-quality Japanese animation image. Preserve the original lighting, format, and colors.

1. Ask ChatGPT to analyze the image Upload the image to ChatGPT and ask the LLM to use its image analysis capabilities to list the defects. Ask it to write a text prompt, once analyzed, to fix them. Prompt: Analyze this image to identify all defects, artifacts, compression errors, and poor quality issues. List them and propose a text prompt to fix them using the GPT Image 2 model, without altering the original colors, proportions, or the artistic style of the lighting and textures. Result: Main issue summary The main problems are compression artifacts, banding, mild posterization, over-smoothing, loss of fine detail, edge halos, slight ringing, and muddy textures. The image should be restored by cleaning artifacts and recovering detail while keeping the original character design, proportions, colors, lighting, and stylized 3D material look unchanged.