What we can capture
Egocentric training data across the full task space — scoped to your environment, modality, and model requirements. All datasets available on request; we scope together before capture begins.
Manipulation
On requestObject grasping, pick-and-place, tabletop assembly, tool use, precision fine-motor tasks.
Annotation layers
Kitchen
On requestFood preparation, chopping, appliance interaction, pouring, utensil use, multi-step cooking tasks.
Annotation layers
Warehouse
On requestPicking, sorting, shelving, bin management, mobile navigation under load, conveyor interaction.
Annotation layers
Manufacturing
On requestAssembly line tasks, part inspection, screw driving, handover operations, precision placement.
Annotation layers
Locomotion
On requestWalking, stair climbing, incline/decline traversal, obstacle avoidance, outdoor terrain.
Annotation layers
Human Activity
On requestDaily life tasks, social interaction, multi-person environments, in-home activities.
Annotation layers
Don’t see what you need? Capture scope is flexible. Tell us your task, environment, and model requirements — we’ll design the dataset with you. Start a scoping conversation →