Everyone can build worlds.
The next internet is three dimensional. We are building the tools that let anyone create for it, not just the few who spent a decade learning how.
Remote first, with a studio in San Francisco
Making a 3D model used to take days of specialist work. That meant whole industries, from small game studios to local shops, were locked out of the most expressive medium we have. Dimensa exists to turn imagination into objects you can hold, light, animate and share, in the time it takes to describe them.
Born from a very long asset list.
Dimensa started in 2024 when our founders were shipping a small indie game. The code was done. The art was not. Four hundred props stood between them and launch, and every AI tool they tried produced blobs that took longer to fix than to model by hand.
So they built a better one. Prism-1 was a research prototype trained on a small, fully licensed library. It made clean, usable meshes, and word spread fast through indie dev communities. Today, Prism-2 powers a full studio used by game teams, retailers, filmmakers and teachers.
We are still small, still shipping every week, and still obsessed with the moment someone types an idea and watches it become real.
Spring 2024
The asset list
Two developers, one game and four hundred missing props.
Winter 2024
Prism-1 research preview
The first model trained only on licensed and commissioned assets.
Fall 2026
Prism-2 and the waitlist
Rigging, retopology and commerce viewers come together in one studio, and the early access waitlist opens.
Q1 2027
Public launch
Open to everyone, with waitlist members invited first.
Principles we actually use.
Artists first
Our models are trained on licensed work and contributing artists share in the revenue. AI should expand creative careers, not erase them.
Useful over flashy
A great render that breaks in an engine is worthless. We measure ourselves by assets that ship in real products.
Craft in the details
Pivots on the ground, sensible UVs, names that make sense. The small things are what make a tool feel professional.
Ship, listen, repeat
We release every week and read every piece of feedback. The beta community shapes our roadmap directly.
Meet Dimi.
Dimi is our in house character and our toughest test case. Every new model, material and rigging feature has to work on Dimi before it ships to you.
- 149k triangles of smooth, beveled parts with no low poly shortcuts
- 36 joint rig with articulated fingers, a blinking visor and a springy antenna
- Seven motions keyed at 60 frames per second for fluid, natural movement
Rendered in Blender CyclesWe're building the founding team.
Researchers, graphics engineers, technical artists and designers who want to make 3D creation as easy as writing a sentence. Remote first, with a studio in San Francisco.
Logos and brand assets
Please use the mark on dark or gradient backgrounds where possible. For interviews and press enquiries, email press@dimensa.ai.
