# lerobot Open-source Python library and tools for machine learning with real-world robots. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Record, train, and deploy from one place with Strands Agents, LeRobot, and Hugging Face Storage Buckets DevFeed: [Record, train, and deploy from one place with Strands Agents, LeRobot, and Hugging Face Storage Buckets]() Original publisher: [Read original article]() Author: Sundar Raghavan; Steven Palma; Cagatay Cali; Arron Bailiss; Yin Song Published: 2026-08-13T17:16:04Z Content type: article Language: en Sources: [Hugging Face - Blog]() Topics: [lerobot](), [hugging face](), [cloud-infrastructure](), [dataset](), [SDKs](), [Command-line interface](), [Open Source]() Tags: [agents](), [apache](), [aws](), [cli](), [code](), [data](), [datasets](), [hub](), [hugging-face](), [lerobot](), [models](), [open](), [open-source](), [robots](), [storage](), [strands-agents](), [train](), [xet]() ### AI overview This article describes a continuous robotics data loop using Strands Agents, LeRobot, Hugging Face Hub, and Hugging Face Storage Buckets. It covers recording demonstrations, collecting episodes, training policies on growing datasets, deploying checkpoints, and using mutable Xet-backed storage to reduce repeated data transfers. ### Source excerpt Record, train, and deploy from one place with Strands Agents, LeRobot, and Hugging Face Storage Buckets You have an agent that can already record a demonstration and push it to the Hugging Face Hub. Now you want to run that loop continuously: collect episodes through the day, train a policy on the growing dataset, deploy it, and pull the next batch back to improve it. Run that loop once and every piece works. Run it every day and you start paying for the same byte transfers over and over. ## Grabette: an open system to record robot-manipulation data DevFeed: [Grabette: an open system to record robot-manipulation data]() Original publisher: [Read original article]() Author: Steve Nguyen; Claire Houziel; Gaelle Lannuzel; Simon Le Goff; Jeremy Laville; Étienne Published: 2026-07-21T00:00:00Z Content type: article Language: en Sources: [Hugging Face - Blog]() Topics: [Robotics](), [datasets](), [lerobot](), [hugging face](), [Open Source]() Tags: [ai](), [community](), [data](), [datasets](), [devices](), [hugging-face](), [lerobot](), [open](), [open-source](), [robot-manipulation](), [robotics](), [video]() ### AI overview Grabette is an open, low-cost handheld system for recording robot-manipulation demonstrations without requiring a robot, laboratory, or teleoperation rig. It uses a human-operated gripper and cameras to capture demonstrations and produce robot-ready datasets, with browser-based processing and integration with LeRobot and the Hugging Face Hub. ### Source excerpt We're on a journey to advance and democratize artificial intelligence through open source and open science. ## LeRobot v0.6.0: Imagine, Evaluate, Improve DevFeed: [LeRobot v0.6.0: Imagine, Evaluate, Improve]() Original publisher: [Read original article]() Author: Steven Palma; Pepijn Kooijmans; Caroline Pascal; Khalil Meftah; Maxime Ellerbach; Martino Russi; Nikodem Bartnik; Nicolas Rabault; Thomas Wolf Published: 2026-07-07T00:00:00Z Content type: article Language: en Sources: [Hugging Face - Blog]() Topics: [lerobot](), [World models](), [Robotics](), [vlm](), [Cosmos](), [foundation-models](), [datasets](), [Open Source]() Tags: [artificial-intelligence](), [cli](), [cosmos](), [datasets](), [gpu](), [lerobot](), [models](), [open-source](), [robotics](), [science](), [training](), [vlm](), [world-model]() ### AI overview LeRobot v0.6.0 adds world-model policies, new vision-language-action models, reward-model APIs, simulation benchmarks, human-in-the-loop CLI corrections, FSDP and cloud training. It also expands dataset capabilities with depth support, automated language annotation, custom video encoding, and faster data loading. ### Source excerpt We're on a journey to advance and democratize artificial intelligence through open source and open science. ## LeRobot v0.5.0: Scaling Every Dimension DevFeed: [LeRobot v0.5.0: Scaling Every Dimension]() Original publisher: [Read original article]() Author: Steven Palma; Pepijn Kooijmans; Jade Choghari; Caroline Pascal; Khalil Meftah; Martino Russi; Nicolas Rabault; Michel Aractingi; Virgile BATTO; Thomas Wolf Published: 2026-03-09T00:00:00Z Content type: release Language: en Sources: [Hugging Face - Blog]() Topics: [lerobot](), [Robotics](), [Unitree](), [Hardware](), [Open Source](), [Artificial Intelligence]() Tags: [artificial-intelligence](), [hardware](), [lerobot](), [open-source](), [python](), [release](), [robotics](), [robots](), [unitree]() ### AI overview LeRobot v0.5.0 expands support for humanoid, mobile, and dual-arm robotics, led by comprehensive Unitree G1 integration. The release also adds new policies, responsive inference, streaming video encoding, simulation-environment loading through EnvHub, IsaacLab-Arena integration, and a modernization of the Python and Transformers stack. ### Source excerpt We're on a journey to advance and democratize artificial intelligence through open source and open science. ## Building a Healthcare Robot from Simulation to Deployment with NVIDIA Isaac DevFeed: [Building a Healthcare Robot from Simulation to Deployment with NVIDIA Isaac]() Original publisher: [Read original article]() Author: Steven Palma; Andres Diaz-Pinto Published: 2025-10-29T00:00:00Z Content type: tutorial Language: en Sources: [Hugging Face - Blog]() Topics: [Isaac for Healthcare](), [Robotics](), [lerobot](), [synthetic-data](), [Training AI Models](), [Inference](), [model-deployment]() Tags: [artificial-intelligence](), [autonomous](), [datasets](), [deployment](), [fine-tuning](), [healthcare](), [inference](), [isaac](), [lerobot](), [nvidia](), [robotics](), [simulation](), [synthetic-data](), [tutorial]() ### AI overview This hands-on guide presents an end-to-end workflow for building an autonomous surgical assistant robot with NVIDIA Isaac for Healthcare. It covers collecting real and synthetic data with LeRobot, fine-tuning GR00T N1.5, evaluating in IsaacLab, and deploying real-time inference on physical hardware. ### Source excerpt We're on a journey to advance and democratize artificial intelligence through open source and open science. ## How to Build a Healthcare Robot from Simulation to Deployment with NVIDIA Isaac for Healthcare DevFeed: [How to Build a Healthcare Robot from Simulation to Deployment with NVIDIA Isaac for Healthcare]() Original publisher: [Read original article]() Author: Asawaree Published: 2025-10-28T20:42:35Z Content type: tutorial Language: en Sources: [Hugging Face - Blog]() Topics: [Isaac for Healthcare](), [Robotics](), [Simulation and Design](), [synthetic-data](), [datasets](), [Training AI Models](), [lerobot](), [model-deployment](), [Inference](), [Medical imaging]() Tags: [ai](), [data](), [datasets](), [deployment](), [healthcare](), [how-to](), [inference](), [isaac-for-healthcare](), [nvidia](), [robotics](), [simulation](), [synthetic-data](), [training](), [tutorial]() ### AI overview This hands-on guide presents NVIDIA Isaac for Healthcare v0.4 and its SO-ARM starter workflow for building autonomous surgical-assistance robots. It covers collecting real and synthetic data with SO-ARM and LeRobot, post-training GR00T N1.5, evaluation in Isaac Lab, and deployment with real-time inference on physical hardware. ### Source excerpt A hands-on guide to collecting data, training policies, and deploying autonomous medical robotics workflows on real hardware Simulation has been a cornerstone in medical imaging to address the data gap. However, in healthcare robotics until now, it's often been too slow, siloed, or difficult to translate into real-world systems. That's now changing. ## LeRobot v0.4.0: Supercharging OSS Robot Learning DevFeed: [LeRobot v0.4.0: Supercharging OSS Robot Learning]() Original publisher: [Read original article]() Author: Steven Palma; Michel Aractingi; Pepijn Kooijmans; Caroline Pascal; Jade Choghari; Francesco Capuano; Adil Zouitine; Martino Russi; Thomas Wolf Published: 2025-10-24T00:00:00Z Content type: release Language: en Sources: [Hugging Face - Blog]() Topics: [lerobot](), [Robotics](), [datasets](), [parquet](), [Streaming](), [Command-line interface](), [hugging face](), [Benchmark](), [Open Source]() Tags: [cli](), [datasets](), [hugging-face](), [lerobot](), [oss](), [parquet](), [robotics](), [simulation](), [streaming]() ### AI overview LeRobot v0.4.0 upgrades open-source robot learning with LeRobotDataset v3.0, supporting chunked episodes, large-scale datasets, efficient video storage and streaming, unified Parquet metadata, faster loading, and dataset editing utilities. The release also adds LIBERO and Meta-World simulation support, simplified multi-GPU training, new VLA models, a plugin system for hardware integration, and a Hugging Face Robot Learning Course. ### Source excerpt We're on a journey to advance and democratize artificial intelligence through open source and open science. ## \`LeRobotDataset:v3.0\`: Bringing large-scale datasets to \`lerobot\` DevFeed: [\`LeRobotDataset:v3.0\`: Bringing large-scale datasets to \`lerobot\`]() Original publisher: [Read original article]() Author: Francesco Capuano; Michel Aractingi; Quentin Lhoest; Caroline Pascal; Pepijn Kooijmans; Jade Choghari; Remi Cadene; Simon Alibert; Adil Zouitine; Martino Russi Published: 2025-09-16T00:00:00Z Content type: article Language: en Sources: [Hugging Face - Blog]() Topics: [lerobot](), [datasets](), [hugging face](), [PyTorch](), [Streaming]() Tags: [datasets](), [hugging-face](), [lerobot](), [open-source](), [pytorch](), [robotics](), [streaming]() ### AI overview LeRobotDataset v3.0 introduces a standardized, extensible format for multimodal robotics data, including sensorimotor readings, camera feeds, teleoperation status, and collection metadata. It integrates with the Hugging Face and PyTorch ecosystems and adds streaming support so large datasets can be processed without downloading them in full. ### Source excerpt We're on a journey to advance and democratize artificial intelligence through open source and open science. ## Post-Training Isaac GR00T N1.5 for LeRobot SO-101 Arm DevFeed: [Post-Training Isaac GR00T N1.5 for LeRobot SO-101 Arm]() Original publisher: [Read original article]() Author: You Liang Tan; Fengyuan Hu; Oyindamola Omotuyi; Oluwaseun Doherty; Chitoku Yato; Shane Reetz Published: 2025-06-11T18:27:22Z Content type: tutorial Language: en Sources: [Hugging Face - Blog]() Topics: [Isaac](), [post-training](), [lerobot](), [Fine-tuning](), [foundation-models](), [datasets](), [multimodal](), [Open Source](), [Python](), [Terminal]() Tags: [data](), [fine-tuning](), [foundation-models](), [hardware](), [isaac](), [lerobot](), [open-source](), [post-training](), [python](), [robots](), [training](), [tutorial]() ### AI overview This tutorial demonstrates how to post-train, or fine-tune, NVIDIA Isaac GR00T N1.5 for a LeRobot SO-101 robot arm using teleoperation data. It covers dataset preparation, GR00T compatibility, installation prerequisites, and running the fine-tuning workflow, with a default requirement of approximately 25 GB of VRAM. ### Source excerpt NVIDIA Isaac GR00T (Generalist Robot 00 Technology) is a research and development platform for building robot foundation models and data pipelines, designed to accelerate the creation of intelligent, adaptable robots. Today, we announced the availability of Isaac GR00T N1.5, the first major update to Isaac GR00T N1, the world's first open foundation model for generalized humanoid robot reasoning and skills. ## LeRobot Community Datasets: The "ImageNet" of Robotics -- When and How? DevFeed: [LeRobot Community Datasets: The "ImageNet" of Robotics -- When and How?]() Original publisher: [Read original article]() Author: Dana Aubakirova; Alexandre Chapin; Mustafa Shukor; Marina Barannikov; Ville Kuosmanen; Remi Cadene; Pedro Cuenca Published: 2025-05-11T00:00:00Z Content type: article Language: en Sources: [Hugging Face - Blog]() Topics: [lerobot](), [datasets](), [Robotics](), [ImageNet](), [Open Source](), [Artificial Intelligence]() Tags: [artificial-intelligence](), [best-practices](), [building](), [collection](), [community](), [data](), [datasets](), [hub](), [lerobot](), [model](), [models](), [open](), [open-source](), [robotics](), [robots](), [science](), [training](), [vision]() ### AI overview This article presents community-contributed LeRobot datasets as a potential "ImageNet" for robotics. It explains that robot generalization across physical, visual, and semantic conditions is constrained by the availability of diverse data, and argues that co-training on heterogeneous datasets can help Vision-Language-Action models adapt across tasks, environments, and robot embodiments. ### Source excerpt We're on a journey to advance and democratize artificial intelligence through open source and open science. ## Hugging Face to sell open-source robots thanks to Pollen Robotics acquisition 🤖 DevFeed: [Hugging Face to sell open-source robots thanks to Pollen Robotics acquisition 🤖]() Original publisher: [Read original article]() Author: Thomas Wolf; Clem 🤗; Matthieu Lapeyre Published: 2025-04-14T00:00:00Z Content type: news Language: en Sources: [Hugging Face - Blog]() Topics: [hugging face](), [Open Source](), [Robotics](), [lerobot](), [reachy](), [Machine learning]() Tags: [acquisition](), [announcement](), [artificial-intelligence](), [community](), [datasets](), [discord](), [education](), [experiments](), [github](), [hardware](), [hugging-face](), [lerobot](), [libraries](), [models](), [open-source](), [reachy](), [research](), [robotics](), [robots](), [youtube]() ### AI overview Hugging Face announces the acquisition of Pollen Robotics to expand its open-source robotics efforts. The article highlights LeRobot, the Hugging Face Hub, and Reachy 2 as part of a vision for accessible robotics research, education, and embodied AI experiments. ### Source excerpt We're on a journey to advance and democratize artificial intelligence through open source and open science. ## LeRobot goes to driving school: World's largest open-source self-driving dataset DevFeed: [LeRobot goes to driving school: World's largest open-source self-driving dataset]() Original publisher: [Read original article]() Author: Harsimrat Sandhawalia; Remi Cadene Published: 2025-03-11T00:00:00Z Content type: article Language: en Sources: [Hugging Face - Blog]() Topics: [lerobot](), [datasets](), [Robotics](), [Artificial Intelligence](), [Open Source]() Tags: [ai](), [artificial-intelligence](), [community](), [datasets](), [lerobot](), [open-source](), [robotics]() ### AI overview LeRobot Goes to Driving School introduces Learning to Drive (L2D), described as the world's largest open-source multimodal self-driving dataset. It contains more than 90 terabytes of data and over 5,000 hours of driving from 30 German cities, including camera footage, vehicle state, driving actions, environmental metadata, natural-language instructions, and future waypoints. The dataset supports training end-to-end automotive and robotics AI models through LeRobot's training pipeline. ### Source excerpt We're on a journey to advance and democratize artificial intelligence through open source and open science. ## π0 and π0-FAST: Vision-Language-Action Models for General Robot Control DevFeed: [π0 and π0-FAST: Vision-Language-Action Models for General Robot Control]() Original publisher: [Read original article]() Author: Dana Aubakirova; Pablo Montalvo; Mustafa Shukor; Remi Cadene Published: 2025-02-04T00:00:00Z Content type: article Language: en Sources: [Hugging Face - Blog]() Topics: [foundation-models](), [Artificial Intelligence](), [lerobot](), [Machine learning](), [datasets](), [multimodal](), [data](), [post-training](), [PyTorch](), [sim-to-real transfer robotics]() Tags: [ai](), [artificial-intelligence](), [community](), [datasets](), [foundation-models](), [lerobot](), [models](), [multimodal](), [post-training](), [pytorch](), [research](), [robotics](), [robots](), [training](), [vision]() ### AI overview The article introduces π0 and π0-FAST, prototype Vision-Language-Action models and learning frameworks for generalist robot control. π0 uses large-scale pretraining and flow matching-based action generation, trained on data from seven robotic platforms and 68 tasks to support dexterous manipulation across different robot embodiments. ### Source excerpt We're on a journey to advance and democratize artificial intelligence through open source and open science.