CynLr Launches Intelligence Platform, Helps Robots Learn in Seconds

Bengaluru-based robotics startup CynLr has unveiled the ‘Object Intelligence’ platform (OI) to help robots learn unknown objects “on the fly”. The company launched the product following five years of research and development.

The platform enables robots to learn, adapt and manipulate unseen objects within 10–15 seconds, without retraining, the company said. The system targets manufacturing and automation users who are constrained by current physical AI methods.

CynLr said the platform can run on industrial robotic arms, multi-arm systems, and humanoid robots. The company said it is the first commercial-ready Object Intelligence platform developed in India.

At the centre of the platform is an OI-enabled vision system called CLX.

The system perceives new objects, builds geometry, and grasps models in real-time. The startup also said this makes it possible for the system to handle transparent, reflective, and irregular objects. These include glass bottles with plastic covers, metallic parts, and electronics.

The company said the new OI platform will support assembly line operations and maintenance tasks in semiconductor fabrication facilities. It stated that switching between trained tasks occurs instantly during deployment.

Configuring tasks within the same setup usually takes about an hour, the company said.

Training for new tasks can take between one week and three months. “Object Intelligence is where it begins,” Gokul NA, Founder at CynLr, remarked at the launch. “The next fifty years will be about cognition,” he said.

Unlike pre-programmed systems, CynLr said its robots learn continuously through interaction. Every grasp becomes a learning event and improves performance in real time.

“The bottleneck for physical intelligence today is sensing,” Founder, GTM at CynLr, Nikhil Ramaswamy said. “Robots need their own eyes, touch and feel to learn and map the world,” he said.

CynLr counts luxury automobile brands, semiconductor fabrication facilities and automation firms as its customers. Examples include automobile manufacturer General Motors and Japanese automotive components manufacturer, DENSO.

The company currently offers two products, CyRo and CyNoid. It also plans to launch an open hardware platform, Mantroid, in 2028. Mantroid will allow customers to design custom robot form factors for specific industrial needs. The company said it is seeking a new funding round to scale production.

“If we can move factories from hardware-defined lines to software-defined machines, we can make manufacturing flexible enough to change products without rebuilding the factory,” Gokul added.

It aims to produce one robot per day by 2028 and expand its global supply chain.

The company also plans to conduct technical demonstrations that showcase real-time perception and learning behaviour.