Feasibility Record

Every retargeted trajectory is replayed in simulation before delivery. Most suppliers use that step as a filter and ship what survives. MOVAS keeps both halves: the trajectories that passed, and a structured record of the ones that did not — which end effector, which clip, which reason.

Rejection reasons are drawn from a closed vocabulary: aperture, clearance, geometry, grasp, scale, deformable. Across five reference end effectors, one clip produces up to five outcomes, so the record grows five times faster than the corpus itself.

The retargeting algorithms are public methods and we make no proprietary claim on them. The accumulated record of what proved infeasible, across embodiments and customers, is not reproducible from public sources — it only exists because the replays were run and kept.