DO-MODE

Inspiration

DO-MODE was inspired by the idea of expressing emotion through pure motion.
Instead of relying on dialogue, I wanted to show a character’s inner state through dynamic movement, color shifts, and cinematic framing.
As an AI animation creator, I’m fascinated by how generative tools can extend human direction rather than replace it.

What I Learned

Throughout the production, I explored how different AI systems interpret motion, lighting, and style.
I learned how to guide AI outputs using detailed storyboards and clear visual references, ensuring the final film stayed consistent with my artistic intention.
This project strengthened my understanding of blending AI motion generation with anime-style visual storytelling.

How the Project Was Built

I started by building the world and overall visual tone using Midjourney and Nanobanana, creating the aesthetic foundation for the short.
After that, I developed a complete storyboard, focusing on anime cinematography, pacing, and movement.

The primary animation was generated with SORA 2, chosen for its coherent motion and cinematic transitions.
I used DomoAI and Vidu selectively for supplemental cuts, action boosts, and specific motions that required additional control or variation.

All cuts were then assembled and refined entirely in Premiere Pro, where I adjusted pacing, enhanced colors, added transitions, and ensured a unified visual style across the entire project.

Challenges

The main challenge was maintaining stylistic consistency while mixing outputs from multiple AI models.
Fast or complex motion sometimes caused distortions, requiring multiple regeneration passes and careful shot selection.
Unifying color, timing, and motion between different systems also demanded precise editing and visual judgment.

Another challenge was balancing AI-generated movement with intentional cinematic direction, ensuring the action remained expressive rather than chaotic.

Outcome

DO-MODE became a stylized AI-animated short that merges bold motion, rich color design, and human-driven direction.
This project represents my ongoing exploration of how AI can expand the language of animation.
Through DO-MODE, I created a workflow that blends generative tools with the creative decisions only a human director can provide.

Built With

• adobe
• domoai
• elevenlabs
• freepik
• midjourney
• nanobanana
• nijivoice
• premirepro
• seedream
• sora-2
• vidu