What we're building
Japan and Korea are already there. China is months behind. Italy, Germany, Spain are next.
In rural Japan today, entire villages sit half-abandoned because there aren't enough people left to maintain the roads, run the clinics, or check on the elderly who stayed behind.
This isn't a forecast. It's a curve that's already locked in. The babies who would have been the workers of 2050 were not born. No policy can reverse it now.
"Humanoid robots are the only realistic answer."
The hardware is arriving. The models are catching up. The technical question is no longer can we build them.
The question is who gets to teach them — and whether the people who most need robots will ever get a seat at that table.
Humanoids can't stay locked behind $50,000 price tags. They have to be cheap enough, open enough, and teachable enough that society itself folds its knowledge into them: step by step, like origami.
The use case that changes everything
Not a video call. Not a voice message.
You open your laptop. You navigate her house through Origami's eyes. You move to the kitchen, boil the water, bring her a cup, and sit across the table while she drinks it.
Origami can display your face. Not a speaker on a stick — a physical presence that looks like you, moves when you move, and lets you be there when you can't be there. You can hold her hand. She can see you smile.
This is what we mean by robotics belonging to everyone. Not just access to robots. Access to each other.
The Three Pillars
Pillar 1
Simulation as a Service + Robot as a Service. Anyone can open a browser, meet a robot, and start teaching it — no hardware, near-zero cost. Together these two services collect natural human–robot interaction data from outside the lab — from teachers, caregivers, small businesses, and students.
Read more →Pillar 2
An open hardware and software humanoid under $5,000, standing no taller than 1.2m, weighing under 15kg. Every schematic, every model, every line of code: open. Built for schools, hobbyists, and anyone who's been priced out of the conversation so far. What you build in simulation transfers directly to the physical robot.
Meet Origami v1.0 →Pillar 3
Research into how an embodied agent stays safe in situations it was never trained for. Pillars 1 and 2 aren't separate from this work — they're how the research gets done. Every simulation session and real-robot interaction generates the diverse, real-world data that makes the question answerable.
Read our research →Currently building. Early access waitlist opening soon.
Talk to Rafa →