Geppetto tracks 1035 robot training datasets including Open X-Embodiment, DROID, and AgiBot World, each with structured metadata on trajectory count, hours of demonstration, robot platforms, license, commercial use, access model, and file format. Every dataset connects to the robots it trains and the professions it affects.

DatasetYearCategoryTrajectoriesHoursPlatformsLicenseCommercialAccessFormat
NavWareSet2026research-lab——mobile robotnot statedNoopen—
STONE Dataset2026research-lab——off-road robot, surround-view multi-modalnot statedNoopen—
Wall-Following Robot Navigation Data2026research-lab——mobile robot ultrasound sensorsnot statedNoopen—
DARP2026research-lab——dual-arm RGB-D-IR rignot statedNoopen—
2026-challenge-demos2026synthetic20,0001,950JoyLoMITYesopen—
FineART2026real-robot40,5431,718Trossen Stationary AICC-BY-4.0Yesopen—
PhysicalAI Robotics GR00T Teleop G12025real-robot——Unitree G1CC BY 4.0Yesopenparquet, hdf5, mp4
Human2LocoMan — Loco-Manipulation Dataset2025real-robot——LocoManMITYesopenhdf5, mp4
RealSource-World — Dual-Arm Humanoid Manipulation Dataset2025real-robot——RS-02 dual-arm humanoidCC BY-NC-SA 4.0 — non-commercial use onlyNogatedparquet, mp4
DexWild — In-the-Wild Dexterous Manipulation Dataset2025human-video——6 DoF dexterous handMITYesopenimagefolder, parquet
EgoVerse — Egocentric Robot Training Dataset: 8.5M Frames, 74 Countries2025real-robot—1,213aria_bimanual, aria_left_arm, aria_right_armCC BY-SA 4.0Yesopenzarr, lerobot, hdf5
AgiBot-Alpha — AgiBot World Dataset2025real-robot—1,979AgiBot A2, AgiBot X1AgiBot World Open Academic License — commercial use with restrictionsYesopenzarr, mp4
LeRobot Community Datasets v32025real-robot——SO-100 arm, SO-101 arm, Franka PandaApache 2.0Yesopenparquet, mp4, lerobot
OXE-AugE2025synthetic4,440,000—Franka Emika Panda, UR5e, xArm7CC-BY-4.0Yesopen—
RFUAV2025research-lab——USRP radio frequency recordings of 37 UAVsnot statedNoopen—
Bench-RNR2025research-lab——infrastructure LiDARs, repetitive and non-repetitive scanningnot statedNoopen—
CoInfra2025research-lab——8-node roadside lidar and camera system at roundaboutnot statedNoopen—
R-LiViT2025research-lab——roadside lidar, RGB and thermal camerasnot statedNoopen—
REHEARSE-3D2025research-lab——LiDAR and radar in emulated rainnot statedNoopen—
DISC2025research-lab——simulated vehiclenot statedNoopen—
SimCoachCorpus2025research-lab——driving simulator, race tracknot statedNoopen—
AevaScenes2025research-lab——Aeva FMCW lidar vehiclenot statedNoopen—
AGC-Drive2025research-lab——vehicles plus UAVnot statedNoopen—
CATS-V2V2025research-lab——two vehicles with lidar and multi-view camerasnot statedNoopen—
Dur360BEV2025research-lab——car with spherical camera, 128-channel lidar, RTK GNSS/INSnot statedNoopen—
EMT2025research-lab——dash camera, UAEnot statedNoopen—
MS2 (Multi-Spectral Stereo)2025research-lab——stereo RGB, NIR, thermal, LiDARnot statedNoopen—
NavINST2025research-lab——multi-sensor road test vehiclenot statedNoopen—
Odyssey2025research-lab——automotive LiDAR-inertial rignot statedNoopen—
OnSiteVRU2025research-lab——vulnerable road user trajectoriesnot statedNoopen—
RASMD2025research-lab——RGB and SWIR camerasnot statedNoopen—
RISEE2025research-lab——naturalistic trajectories with risk ratingsnot statedNoopen—
RoadSens-4M2025research-lab——smartphone sensors and cameranot statedNoopen—
RoRaTrack2025research-lab——Dallara AV-21 cameranot statedNoopen—
STU (Spotting the Unexpected)2025research-lab——lidar and cameranot statedNoopen—
TAD-E2E2025research-lab——not stated on page openednot statedNoopen—
UrbanIng-V2X2025research-lab——multiple vehicles and infrastructure across intersectionsnot statedNoopen—
UrbanV2X2025research-lab——vehicle and infrastructure sensors, C-V2Xnot statedNoopen—
V2X-ReaLO2025research-lab——V2X-Real vehicles and infrastructure, ROS bagsnot statedNoopen—
Waymo internal driving dataset (Scaling Laws of Motion Forecasting and Planning)2025research-lab——not statednot statedNoopen—
DexUMI2025human-video——wearable hand exoskeletonnot statedNoopen—
Aria Gen 2 Pilot Dataset2025human-video——Project Aria Gen 2 glassesProject Aria dataset terms (per atlas)Noopen—
Aria STDR2025human-video——Project Aria glassesnot statedNoopen—
Digital Twin Catalog (DTC)2025human-video——Project Aria glasses and DSLRdataset licence agreementNoopen—
ECBench2025human-video——egocentric videosnot statedNoopen—
Ego-HOIBench2025human-video——Egocentric images with hand-verb-object tripletsnot statedNoopen—
EgoBlind2025human-video——cameras used by blind peoplenot statedNoopen—
EgoCross2025human-video——egocentric video from surgery, industry, extreme sports, animal viewsnot statedNoopen—
EgoEdit (EgoEditData)2025human-video——egocentric videosnot statedNogated—
EgoEvGesture2025human-video——egocentric event cameranot statedNoopen—
EgoExo-Con2025human-video——synchronised ego-exo video pairsnot statedNoopen—
EgoGazeVQA2025human-video——Egocentric daily-life video with gazenot statedNoopen—
EgoIllusion2025human-video——egocentric videosnot statedNoopen—
EgoNight2025human-video——egocentric video, day and night pairsnot statedNoopen—
EgoNormia2025human-video——egocentric videos of human interactionsnot statedNoopen—
EgoPrivacy2025human-video——existing egocentric videosnot statedNoopen—
EgoRe-5M (EgoThinker)2025human-video——Egocentric video clips from existing sourcesnot statedNoopen—
EgoSocial2025human-video——Egocentric social interaction videonot statedNoopen—
EgoTempo2025human-video——egocentric videosnot statedNoopen—
EgoTextVQA2025human-video——ego-view videos, driving and house-keepingnot statedNoopen—
EgoToM2025human-video——Ego4D videosnot statedNoopen—
EgoVLA training mix and Ego Humanoid Manipulation Benchmark2025human-video——egocentric human videos; sim benchmark uses Unitree H1 with Inspire handsMIT (Hub EgoVLA/EgoVLA-Humanoid-Sim)Yesopen—
EMMA2025human-video——Project Aria glassesnot statedNoopen—
EOC-Bench2025human-video——Egocentric videonot statedNoopen—
EvHand-FPV2025human-video——Single event camera, first-person viewnot statedNoopen—
ExAct2025human-video——skilled-activity ego and exo videosnot statedNoopen—
GuideDog2025human-video——pedestrian first-person imagesnot statedNoopen—
LSC-ADL2025human-video——lifelog wearable camera images (LSC)not statedNoopen—
TASTE-Rob2025human-video——fixed-view ego-centric cameranot statedNoopen—
TeleEgo2025human-video——wearable camera with audionot statedNoopen—
UniHand (Being-H0)2025human-video——mixed sources: motion capture, VR headsets and RGB-only videosnot statedNoopen—
DG16M2025research-lab——dual-arm graspsnot statedNoopen—
Dexonomy2025research-lab——dexterous hands, 31 GRASP taxonomy typesnot statedNoopen—
Learning Gentle Grasping dataset2025research-lab——multi-fingered hand with vision, sound and touchnot statedNoopen—
ScaleADFG2025research-lab——3D assets with dexterous functional graspsnot statedNoopen—
UniFucGrasp2025research-lab——multiple dexterous hand typesnot statedNoopen—
SpikeGrasp2025research-lab——stereo spike camerasnot statedNoopen—
TD-TOG2025research-lab——real RGB-D scenes with affordances and planar graspsnot statedNoopen—
GraspFactory2025research-lab——Franka Panda and Robotiq 2F-85 gripperscustom licence fileNoopen—
LIBERO-Plus2025synthetic—31.08franka_pandaMITYesopen—
Humanoid Everyday2025real-robot—27.8unitree_humanoidApache-2.0Yesopen—
AIRoA MoMa Dataset2025real-robot—94—customNoopen—
U2UData+2025synthetic—120droneApache-2.0Yesopen—
RoboArena2025real-robot——franka_panda, parallel_jawMITYesopen—
EgoDex2025human-video—829humanresearch-onlyNoopen—
UAV-Flow-Sim2025synthetic——droneunknownNoopen—
UAV-Flow2025human-video——droneunknownNoopen—
CognitiveDrone Dataset2025synthetic——droneunknownNoopen—
VideoPhy-22025research-lab——generated action-centric videoMITYesopen—
1000 Robot Manipulation Tasks2025research-lab——not statedmitYesopen—
X-Capture2025research-lab——portable device: RGB-D, tactile, audio, 3DMITYesopen—
OpenLoong mocap data2025research-lab——human BVH skeleton for OpenLoong humanoidApache 2.0Yesopen—
Unitree Z1 Dual-arm datasets2025real-robot——Unitree Z1 dual arm with Dex1 grippersApache-2.0Yesopen—
FeelAnyForce2025research-lab——not stated on cardCC BY 4.0Yesopen—
VitaSet2025research-lab——RGB vision and tactile sensing via robotic manipulationMITYesopen—
Wanderland2025research-lab——not statedapache-2.0Yesopen—
PhysicalAI-Robotics-GR00T-GR1 (GR1-100)2025real-robot——Fourier GR1-T2 (real), third-person video onlyCC BY 4.0Yesopen—
WorldScore2025research-lab——static and dynamic scenes with camera trajectoriesMITYesopen—
ManiFeel2025synthetic——visuotactile manipulation benchmarkMITYesopen—
Mixed Signals2025research-lab——3 connected vehicles plus roadside unit, 5 lidarsCC BY-NC-SA 4.0Noopen—