New Delhi, Feb. 12 -- Deeptech startup CynLr on Thursday unveiled its commercial-ready Object Intelligence (OI) Platform, a robotics system that enables machines to learn and adapt to previously unseen objects within 10-15 seconds - without offline retraining.

The five-year R&D effort positions the Bengaluru-based company among a small set of global players attempting to move robotics beyond pre-programmed automation toward real-time cognition. CynLr claims this is the first commercially ready Object Intelligence platform to emerge from India.

The platform powers CynLr's robots - CyRo, CyNoid and the upcoming Mantroid - and is form-factor agnostic, meaning it can be deployed across industrial robotic arms, multi-arm systems and humanoid configurations. The company said leading luxury automotive brands and semiconductor automation firms are already evaluating or deploying the system for assembly-line operations and maintenance tasks in semiconductor fabrication units.

At the core of the platform is CLX, an OI-enabled vision system that can perceive an unseen object and break it down into geometry, texture, reflectance and grasp possibilities in real time. The robot then experimentally learns how to manipulate the object through interaction. CynLr said the system can handle transparent, reflective and irregular objects - such as glass bottles wrapped in plastic, metallic automotive parts and electronics components - challenges that typically limit conventional physical AI systems.

Unlike traditional robotics, which relies on months of dataset training and structured environments, CynLr's approach is built on closed-loop sensorimotor learning inspired by neuroscience. The company has collaborated with the Indian Institute of Science (IISc) on research aimed at building perception systems closer to brain-like adaptation.

Founder Gokul NA said the shift marks a transition from programmed automation to cognitive machines. "The last fifty years of robotics were about repetition. The next fifty will be about cognition - machines that observe, reason and adapt," he said, outlining the company's long-term vision of a "Universal Factory" where production lines can switch between products through software without physical retooling.

He said task switching within trained setups is instantaneous, while new task configurations can take from an hour to a few months depending on complexity. In several complex assembly scenarios, the company argues, Object Intelligence enables automation where none previously existed.

Founded in 2019, CynLr is now preparing for its next funding round to scale manufacturing, targeting production of one robot per day by 2028. It has also established an R&D entity in Switzerland and a US business development centre as part of its global expansion strategy

Published by HT Digital Content Services with permission from TechCircle.