Operational Showcase Active • India 🇮🇳

Inspect human demonstration data.

Browse and play multi-angle task sequences across operational factory floors, retail outlets, and household environments. Inspect bimanual state annotations and download schemas.

Inspection Console: Household Cooking Veg Prep
Household Data1080p • 30fps
Speed:

Dataset Scope

Bimanual cucumber peeling and slicing rounds using a claw safety grip on a cutting board.

Core EnvironmentControlled residential kitchen environment
Camera Modalities1x Egocentric (1440x1440 60fps), 3x Exocentric static tripod (1080p 30fps)

Technical Details

Duration Profile16s sequence run
Delivery DeliverablesH.264 MP4 task sequences + structural state annotations (.json / .csv)
Action AnnotationsBimanual task trajectories, object bounding boxes, and action phase flags
Ready to test sample clips in your model?Get evaluation-ready media packages with complete schemas.
Consent-firstVerified commercial release rights
Multi-angle ready1x stereo egocentric rig + 3x exocentric cameras
Rich State LogsActive hand trajectories and bounding boxes

Capture hardware

Stereo camera specification.

The egocentric rig records synchronized visual, motion, audio, and hand signals in a single training-ready capture.

CamerasDual 1080p at 30 fps · Global shutter
Field of view157° × 84° per camera
Stereo baseline63 mm — matched to human IPD
IMU9-DoF at 400 Hz
AudioStereo, dual microphones
Hand detectionEgoHand — real-time, on-device, every frame*
SynchronizationAll four streams on one hardware clock
EncodingH.265 / HEVC, on-device
OutputOne time-aligned .mcap file
StorageMicroSD — 128 GB card included
ConnectivityWi-Fi + USB — live preview, upload, direct transfer
Voice promptsBuilt-in speaker — session and status announcements
MountAdjustable pitch bracket
PowerAny USB power bank — runtime scales with capacity
Weight≈ 200 g, headset only

* Hand detection runs on-device for every frame.

Evaluation-ready

Know exactly what your team receives.

A focused sample package gives technical and procurement teams enough evidence to assess quality, rights, and integration fit.

Curated task clipsRepresentative sequences from the environments and workflows you care about.
Rights documentationClear consent methodology and commercial-use status for buyer review.
Metadata previewExample JSON or CSV schemas with task, object, environment, and QA fields.
Capture recommendationA practical path from evaluation sample to a production collection program.

Test the data before you scale it.

Send the environment, objects, robot task, target hours, and schema. We will scope a focused evaluation sample and collection path.