Cross the reality gap with evidence, not optimism.
A policy that wins in simulation and fails on hardware is worse than no policy. Kyros combines domain randomization, system identification, and transfer evaluation so teams can see risk before a robot moves.
transfer_eval_suitehardware rollout confidence
sdk pipelineeval run
seedscoreexport
physics
policy
transfer
94%
Average sim-to-real success rate across Transfer Verified environments
3.2×
Improvement over default Isaac/MuJoCo randomization presets
<2 hrs
From trained policy to first hardware rollout via the Kyros eval pipeline
Robust
Domain RandomizationKyros Randomizer
Mass, friction, motor dynamics, latency, lighting, textures, camera pose — every nuisance variable is parameterized and randomized per episode. Your policy has to be robust by construction.
#Dynamics#Textures#Friction
Train Task →Calibrated
System IdentificationReal-world fit
Calibrate your simulator against real-world rollouts. Kyros fits sim parameters to measured trajectories so your training distribution matches your robot.
#Calibration#Fidelity
Train Task →Score-based
Transfer EvaluationEval suite
Run any policy through a standardized eval suite that produces a transfer-likelihood score. Stop guessing whether your policy will work outside the lab.
#Likelihood#Standardized
Train Task →Protected
Safety SandboxingSandbox
First deployments happen in a hardware-in-the-loop sandbox with safety envelopes, action clipping, and automatic rollback on failure.
#Hardware-in-Loop#Rollback
Train Task →The robots of tomorrow need training grounds today.
Kyros guarantees physics-accurate model parameters that port natively to your production robot stack.