CLARU
xuejfcc-by-nc-sa-4.0

EgoPro

1,000 hours of synchronized head-and-wrist egocentric video for fine-grained manipulation tasks, with 3D hand pose, body pose, and semantic annotations.

Downloads0
Hours1000

Why This Matters for Physical AI

Synchronized egocentric dual-view video with 3D pose annotations provides a rich foundation for training embodied AI systems to understand and execute fine-grained manipulation tasks from human demonstrations.

Technical Profile

Modalities
rgbhand_pose_3dbody_pose_3dsemantic_annotation
Environment
lab
Task Types
manipulation
Total Hours
1000
Data Format
parquet
Annotation Types
hand_pose_3dbody_pose_3dsemantic_annotation
License
cc-by-nc-sa-4.0
Part of the EgoSuite-Open10K family

Access

View on HuggingFace

Need custom rgb data?

Claru builds purpose-built datasets for lab applications with dense human annotations and quality assurance.

Request a Sample Pack

Related Datasets

EgoSuite-Open10K

A human-centric egocentric video dataset containing 10,000 selected clips organized into 6 SKUs across EgoStandard and EgoPro families for Embodied AI and Physical AI research. The dataset includes egocentric and wrist-view perspectives with annotations for hand pose, body pose, and semantic segmentation across 7 scene groups.

rgbhand_posebody_posesemantic_segmentation
33 downloadsAug 2026apache-2.0

EgoSuite-Open10K

Real-world egocentric human activity dataset with 10,000 hours of video across 7 environment families, captured from head and wrist viewpoints with structured pose and semantic supervision for embodied AI and robot learning.

rgbhand_posebody_posesemantic_annotation
26 downloadsAug 2026

EgoSuite-Open10K

10,000 hours of real-world egocentric human activity data organized into six composable data SKUs with head and wrist viewpoints, designed for embodied AI and world-model research.

rgbhand_posebody_posesemantic_labels
24 downloadsAug 2026

EgoDemo

A 20-50 hour ungated trial pack sampled from the EgoSuite-Open10K repositories, featuring egocentric video demonstrations with human actions and hand/body pose annotations.

rgbhand_pose_3dbody_pose_3dsemantics
0 downloadsAug 2026

EgoStandard

9,000 hours of head-view egocentric video capturing human activities with aligned 3D hand pose, body pose, and semantic annotations across three progressive supervision configs.

rgbhand_pose_3dbody_pose_3dsemantic_annotation
0 downloadsAug 2026cc-by-nc-sa-4.0

VIFailBack (TsFile)

Apache TsFile conversion of the LeRobot v3.0 ALOHA AgileX Piper robot dataset containing time-series and tabular robot state, target, timing, episode/task tags, and debug signals from 4,714 episodes across 98 tasks.

timeseriestabularrgb
null downloadsAug 2026MIT