I’m Yuri Na, the founder and CEO of Twin AI.
Twin AI is a next-generation platform that creates 2D and 3D avatars for the metaverse—digital twins that can represent who we are in both virtual and physical worlds.
I started this company with a simple question: In the era of AI, how will our identity and presence evolve beyond the physical world?
As remote work, virtual spaces, and AI-driven collaboration become part of daily life, people will need authentic digital versions of themselves—avatars that reflect not just their appearance but their individuality, creativity, and professional value.
At Twin AI, we’re building the bridge between reality and the metaverse. Our goal is to empower everyone to express, connect, and work freely across worlds—with technology that makes virtual existence as meaningful as real life.
This is more than character generation. It’s the beginning of a new human-AI era, where identity, creativity, and presence coexist seamlessly.
Goal. Generate avatars that can flex between photoreal “human” and stylized “metaverse” looks while preserving personal identity.
Input. A real portrait (2nd image row) and optional prompts/attributes (age vibe, profession, outfit).
Output. Three levels shown left → right:
Level H (Human-like). Natural skin shading, realistic proportions, subtle textures. Closest to the person’s real look.
Level M (Meta-hybrid). Clean geometry, softened features, slight eye/skin exaggeration, smoother materials.
Level S (Stylized Metaverse). Iconic features, larger eyes, simplified materials, “CG-clean” finish optimized for virtual worlds.
What “more meta” means here. Increasing stylization by:
face/eye proportion shifts, albedo simplification, reduced micro-texture, higher gloss uniformity, color harmony, and toon-friendly lighting. “More human” restores real-world textures, subsurface scattering cues, and true-to-life anatomy.
Controllability. One slider controls the style domain (Human 0.0 ↔ Meta 1.0). We can lock identity with feature anchors (eyes, nose, jaw) and adjust style independently. Export targets: FBX/GLB/VRM with rigging for Unity/Unreal.
Applications. Corporate avatars, education, telepresence, creator economy, and branded IP characters—consistent identity across 2D/3D and platforms.
The images below illustrate the evolution of our development journey over the past few years — from my early coding experiments to building a full-scale 2D and 3D character generation system for the metaverse. Each stage reflects how our technology has continuously improved in image quality, realism, and creative expression, shaping the foundation of Twin AI today.
If you’d like to see more, I’d be happy to share a detailed portfolio upon request.
I’m also open to a short coffee chat or a 30-minute Zoom meeting to discuss our vision and technology further.
📩 Feel free to reach out on LinkedIn — I’d love to connect.
2D & 3D AI Metaverse Character Maker
Original 2D Metaverse Character 3D Metaverse Character
There are many AI cartoonizer programs but my algorithm can create 2D and 3D both at the same time.
It only takes 2 to 3 seconds for 2D and 20 to 30 seconds for 3D (it depends on the image size)
[My 1st algorithm to 2nd algorithm "Quality Improvement"]
applied image enhancement and texture, contrast, symmetry, histogram matching(brightness),
level set(active contour method), boundary(sensitivity, noise adjustment, sharpness), Laplacian, filtering, etc.
It doesn't matter if the image quality is bad or good, my algorithm still works fine.
*I don't support any of the political parties, just used famous people as input images*
Based on my photo(#input->my face photo), it creates my metaverse character automatically and takes 5 seconds for the whole body.
(Higher quality version) takes 20-30 seconds for face, whole body takes less than 50 seconds)
Now I am adding more functions, if you don't like some parts, you can change the shape and styling, body, face, eye, hairstyle, etc.
Metaverse Shopping Mall, trying a new style of clothes
I created a woo-sung Jung (Actor)'s metaverse character.
#Input -> woo-sung Jung's face photo