published index · updated 2026-08-11
Search source-backed records by access, signals, readiness, and adoption. Unknown facts remain visible.
Source-backed datasets and embodied models.
NVIDIA · 2025-03-18
Open foundation model for generalist humanoid robots, focused on whole-body and manipulation behavior from multimodal robot data.
338 downloads
Readiness 76 · verified 2026-08-09
3 linked datasets
Hugging Face · ropedia-ai/xperience-10m · 2026-03-11
A large egocentric multimodal dataset with synchronized video streams, audio, depth, poses, mocap, IMU, and hierarchical language annotations.
10M experiences
Readiness 77 · verified 2026-08-09
5 linked models
Hugging Face LeRobot · 2024-01-26
Compact open vision-language-action policy designed for practical robot fine-tuning and deployment through the LeRobot ecosystem.
73.5K downloads
Readiness 76 · verified 2026-08-09
3 linked datasets
Hugging Face · facebook/ego-1k · 2026-01-29
A synchronized 12-camera egocentric dataset for neural 3D/4D reconstruction, hand-object interaction, and novel-view synthesis.
956 clips
Readiness 74 · verified 2026-08-09
3 linked models
X Square Robot · 2026-05-29
An open 4B VLA pretrained across more than 20 embodiments, with physical-hardware evaluation of pretrained robotic capability.
1.4K downloads
Readiness 76 · verified 2026-08-09
3 linked datasets
Google DeepMind / Open X-Embodiment collaboration · 2023-10-13
A large multi-institution collection spanning 22 robot embodiments and hundreds of skills for cross-robot policy learning.
22 robots
Readiness 77 · verified 2026-08-09
7 linked models
Shanghai AI Laboratory / collaborators · 2025-01-27
A spatially enhanced 4B VLA pretrained on 1.1 million real-robot episodes with explicit geometry-aware representations.
651 downloads
Readiness 76 · verified 2026-08-09
3 linked datasets
AgiBot (Zhiyuan Robotics) · 2026-03-01
AgiBot World 2026 is the latest release of the AgiBot World embodied-AI dataset, succeeding the earlier Alpha/Beta releases. It is collected in real-world commercial, home, and general-purpose environments using the AGIBOT G2 wheeled humanoid with dual 7-DOF force-controlled arms, capturing synchronized RGB(D), LiDAR point clouds, tactile, IMU, and full-body joint states. Released in sequential phases, it provides long-horizon manipulation, navigation, dual-arm coordination, and human-robot collaboration data in LeRobot format (~9.4 TB).
2976.4 hours
Readiness 77 · verified 2026-08-10
5 linked models
starVLA · 2025-10-09
StarVLA: A Lego-like Codebase for Vision-Language-Action Model Developing
3.4K stars
Readiness 69 · verified 2026-08-08
MMLab@HKU and collaborators · 2026-07-05
**RoboDojo** is a sim-and-real evaluation benchmark built to answer a blunt question: how good are generalist manipulation policies really? The answer it reports is sobering — the best of **30 policies evaluated in simulation** reached only **8.80%** success against **76.03%** for human teleoperation. - **60 tasks total**: **42 simulation tasks** (NVIDIA Isaac Sim 5.1 / Isaac Lab 2.3) spanning five capability axes — Generalization (12), Memory (6), Precision (8), Long-Horizon (8), Open-Vocabulary (8) — plus **18 real-world tasks** across three embodiments (ARX X5, Piper, Piper X). - Ships a **training dataset**: **3,500 simulated trajectories** (1,859,602 frames, **20.66 h**) and **1,800 real-world trajectories** (1,611,841 frames, **17.91 h**), all bimanual and recorded at **25 Hz**. - Distributed in several formats — **LeRobot v3.0** (120 GB), LeRobot v2.1 (64 GB), **HDF5** (523 GB), depth HDF5 (~4.5 TB), and real-world data (273 GB). - Real-world rigs use **3 synchronised cameras**: one head camera (Gemini 335L) and **two wrist cameras** (Gemini 305). Evaluation is **10 trials × 18 tasks = 180 real trials** per policy, scored double-blind by three independent evaluators. - **RoboDojo-RealEval** is a cloud-accessible real-robot evaluation service with standardised hardware and scene reset, so a policy can be benchmarked on physical robots remotely without owning any. - Policy integration runs through **XPolicyLab**, which unifies **40+ policies** behind one interface. - Simulation demonstrations come from two sources: **automated trajectory synthesis** composed from motion-planning primitives (grasp, place, handover, insert, open, close, stack, push_up), and **VR teleoperation** for tasks too complex to synthesise. Real-world demos are **leader-follower teleoperation**, 100 per task from four operators. Assembled by a 43-author consortium led by MMLab@HKU, with contributors from UC Berkeley, Tsinghua, Peking University, Stanford and MIT.
5300 episodes
Readiness 77 · verified 2026-08-09
2toinf · 2025-09-25
[ICLR 2026] The offical Implementation of "Soft-Prompted Transformer as Scalable Cross-Embodiment Vision-Language-Action Model"
703 stars
Readiness 69 · verified 2026-08-09
UC Berkeley / MIT / Amazon FAR / CMU / XDOF · 2026-06-25
ABC-130k is an open bimanual robot manipulation dataset released with the paper 'Scalable Behavior Cloning with Open Data, Training, and Evaluation'. It contains roughly 130,919 episodes (43,090 annotated) totaling about 3,598 hours collected on bimanual YAM stations with two 6-DoF arms and parallel-jaw grippers, spanning ~195 tasks such as pick-and-place, folding, sorting, handover, insertion, tool use, and assembly. Each episode includes three camera views (one top plus two wrist), joint states, end-effector poses, gripper aperture, and task/subtask annotations, plus 400 hours of accompanying sim-teleop data.
130919 episodes
Readiness 73 · verified 2026-08-09
openvla · 2024-06-13
OpenVLA: An open-source vision-language-action model for robotic manipulation.
6.8K stars
Readiness 69 · verified 2026-08-09
Galaxea AI · 2025-08-30
The Galaxea Open-World Dataset is a large-scale real-world mobile manipulation collection gathered with a single embodiment, the Galaxea R1 Lite dual-arm mobile manipulator, across home, kitchen, retail, and office environments. It contains roughly 100K demonstration trajectories totaling 500+ hours, paired with fine-grained subtask-level language annotations. Data is released in LeRobot v2.1 format with four RGB camera streams plus proprioceptive state, and underpins the G0 dual-system VLA model.
100000 episodes
Readiness 73 · verified 2026-08-09
4 linked models
IPEC / EO-Robotics · 2025-08-28
A unified embodied model using interleaved vision-text-action pretraining for perception, reasoning, planning, and continuous robot control.
44 downloads
Readiness 76 · verified 2026-08-09
3 linked datasets
Hugging Face · SUZ-tsinghua/egocentric_adjust_bottle · 2026-05-02
A LeRobot-style robotics dataset for an egocentric bottle adjustment task, packaged as Parquet with image, text, and timeseries modalities.
10K-100K rows
Readiness 73 · verified 2026-07-28
4 linked models
Robotics Diffusion Transformer research · 2024-10-10
Diffusion foundation model for bimanual manipulation that uses large-scale robot data to generate action trajectories.
319 downloads
Readiness 76 · verified 2026-08-09
3 linked datasets
Hugging Face · haoyang-li/EgoWorld · 2026-03-30
A compact egocentric bimanual manipulation dataset with world-frame 3D hand poses, MANO meshes, camera trajectories, depth maps, and 40D action/state vectors.
2 episodes
Readiness 63 · verified 2026-07-28
13 linked models
Robbyant · 2026-04-15
A feed-forward 3D foundation model for reconstructing scenes from streaming data
16.3K stars
Readiness 75 · verified 2026-08-09
Hugging Face · zeno-labs/egostation-gopro-pick-and-place-v1 · 2026-05-12
A non-gated CC-BY-NC-4.0 LeRobot-style dataset with GoPro first-person manipulation, hand-world coordinates, and 6DoF trajectory tags.
10K-100K rows
Readiness 57 · verified 2026-07-28
9 linked models
Robbyant · 2026-01-29
[RSS 2026] Causal video-action world model for generalist robot control
1.7K stars
Readiness 75 · verified 2026-08-09
BitRobot · 2026-06-25
HIW-500 (Humanoids In-the-Wild) is a large-scale dataset for whole-body humanoid robot learning in natural home environments, capturing human teleoperation demonstrations on the Unitree G1 across 12 real homes in Southeast Asia. It contains 500+ hours and 23K+ episodes spanning 10+ household tasks with 161 subtask labels (~10 TB). Each episode records synchronized head (stereo RGB) and wrist (RGB + stereo IR) cameras, 29-DoF joint states, end-effector state, IMU, odometry, and language annotations.
23000 episodes
Readiness 77 · verified 2026-08-10
nvidia · 2026-02-25
Official model catalog record.
37.8K downloads
Readiness 76 · verified 2026-07-28
LeRobot (Hugging Face) · 2026-06-22
The shirt-folding dataset from LeRobot's open-source 'Unfolding Robotics: Open-Source Shirt Folding from Data to Deployment' project, which trained a bimanual robot to fold t-shirts at a 90% success rate. Collected via real-world leader-follower teleoperation across 8 bimanual setups optimizing for diversity (25+ t-shirts, 8 backgrounds, varied camera/robot heights), structured as Level 1 (fold a laid-out shirt) and Level 2 (spread a crumpled shirt, fold, set aside). The full dataset has 5,688 episodes (~131h); a curated high-quality subset (lerobot/high_quality_folding) has 1,200 episodes (~30h). Each episode has 3 camera streams (base 480x640, two wrists 720x1280) and 16-dim joint+gripper states/actions, stored in LeRobotDataset v3.0 (video-encoded).
5688 episodes
Readiness 70 · verified 2026-08-10
Showing 24 of 353
Readiness summarizes published evidence and confidence. It does not replace license review, sample inspection, or runtime tests.
Source-backed fields only