So what is Linum?
A two-person research lab training generative video models from scratch.
Technology has completely upended unscripted tv and film. Anyone can shoot on their iPhone, edit on their laptop, and blast it out to the world over the internet. It hasn't happened for movies, because it's just too expensive to craft one. When it comes to animation specifically, you'd expect the barrier to entry to be lower. Everything can be done on just a laptop, but it still costs $3-4M to make an indie like Flow.
The problem lies in the tools. Traditional 3D animation software is functionally rich but semantically poor. You can make anything in theory, but it's really hard to just get started and make something. We think generative video models can fix that.
At Linum, we're building creative controls that animators actually need (consistent characters, intentional camera movement, controllable motion). These can't be bolted on through fine-tuning or complicated workflows. These primitives need to be baked into the model training itself. That's why we build our own models from scratch. It's the first step towards a new set of creative tools that can make animation accessible to everyone.
Backed by Y Combinator
Get in touch
If you're a filmmaker or animator curious about what our models can do, we'd love to hear from you. We're also always interested in talking to anyone with compute or data to offer.






