Chetana
चेतना/ chetanā
One model that perceives, reasons, and acts in a body.
Chetana is the brain — a general embodied-control model that turns intent into coordinated whole-body action. It composes the other systems into a single agent, and it is the work through which we pursue AGI-grade general intelligence in machines.
Flagship · in active development
Compiled, not trained
Control is synthesised in closed form from human motion — no gradient training, no reinforcement-learning rollouts.
Goal-conditioned
Takes an instruction or a goal and plans the whole-body behaviour to reach it.
Body-agnostic
One cortex generalises across chassis — bipeds, arms, mobile platforms.
Toward AGI
The north-star programme: general intelligence that is embodied, grounded, and acts in the physical world.
Today's robots are narrow: each behaviour is trained, tuned, and brittle. Chetana is built the other way around — a single control intelligence that is given a body and a goal, and works out the coordinated action itself.
It is the integration layer for the whole Yantra stack: it sees through Drishti, moves through Gati, manipulates through Hasta, and runs on Kshana at the edge. The skills it draws on live in Kosha.
General intelligence in a body is the aim — an agent that transfers what it knows from one task and one chassis to the next. We are honest that this is a research frontier; Chetana is how we walk toward it, capability by capability.
Next system
Gati गति — Human-like motion.
Walk, balance, recover, gesture — like a person.