Robot tasks with the spec work already done.
Every Kyros task ships with reward functions, success criteria, observation spaces, action constraints, and evaluation harnesses. You write the policy; the training contract is already versioned.
task_registryreward + success criteria
logistics simmobile base
binsroutesfork path
physics
policy
transfer
570+
task specs spanning manipulation, locomotion, perception, and logistics
Typed
observation and action spaces for common robot platforms
Eval
standardized scoring before policy export or hardware rollout
Active
Manipulation180+ tasks
Grasping, pick-and-place, in-hand reorientation, tool use, deformable manipulation, multi-arm coordination.
#Pick & Place#Grasping#In-Hand
Train Task →Verified
Locomotion80+ tasks
Bipedal and quadrupedal walking, stair traversal, recovery from disturbance, terrain adaptation.
#Bipedal#Quadrupedal
Train Task →Active
Perception-Driven120+ tasks
Visual servoing, semantic grasping, dynamic obstacle avoidance, language-conditioned policies.
#Vision#Obstacles#ML
Train Task →Contact-Rich
Assembly & Insertion60+ tasks
Peg-in-hole, screw fastening, cable routing, gear meshing — high-precision contact tasks.
#High-Precision#Peg-in-Hole
Train Task →Active
Logistics90+ tasks
Bin picking, packing, sorting, palletization across SKU diversity and clutter levels.
#Bin Picking#Packing
Train Task →Advanced
Whole-Body Control40+ tasks
Coordinated mobile manipulation, dynamic balance under load, reactive whole-body planning.
#Mobile Manip.#Balance
Train Task →Each task is a complete spec.
No hand-rolled reward shaping. No undocumented success criteria. Every task is reproducible from the registry.
Reward function
Tunable, documented, version controlled. Default rewards have been validated against transfer outcomes.
Success criteria
Explicit pass/fail metrics. No ambiguity over what 'solved' means.
Observation & action spaces
Well-typed observation/action specs that match common robot platforms out of the box.
Evaluation harness
Run any policy against the standard evaluation suite. Get a transfer-likelihood score.
Skip the spec work. Start training.
Get early access to the task registry and standardized evaluation harnesses.