Teach robots by showing, not programming.
proxyrobots is a Physical AI software start-up that lets industrial robotics learn tasks the way humans do - by understanding a demonstration scenario, saving money spent on deterministic programming and planning .

Industrial automation is stuck. Programming a robot for a single task takes weeks, requires specialized engineers, and costs tens of thousands of euros per deployment. The moment the product changes, you start over. That's why most repetitive work in manufacturing and warehousing is still done by hand.

A non-technical operator puts on a VR headset and performs the task naturally. Our Physical AI captures the motion, intent, and context  and transfers it directly to the robot. When the task changes, re-demonstrate in minutes. 
No programmers. 
No consultants. 
No integrators. 
No production downtime.

Use cases: pick-and-place, packaging, machine tending, repetitive assembly, the high-volume, high-variance work that factories/warehouses can't staff and can't afford easily to automate the traditional way.

Foundation models for robotics are hitting the inflection point computer vision reached a decade ago. Hardware is commoditizing (Universal Robots, Agilex, KUKA, etc.). The bottleneck has shifted to software. That's where we live.

Co-founded by Vlad Fatu (https://www.linkedin.com/in/vlad-fatu-697a5143/) and Robert Simion (https://www.linkedin.com/in/robert-simion-b25110133/). 

Accelerated in Innovation Labs (https://www.innovationlabs.ro/) — Romania's leading deep-tech accelerator.

We're talking to investors, robotics hardware partners, and pilot customers. If automation has been on your roadmap but off your budget, let's talk!

📩 contact@proxyrobots.com
🌐 proxyrobots.com

proxyrobots

Org chart

Teams

Offices

proxyrobots

Org chart

Teams

Offices

Related companies