
MANUS data gloves capture every finger, joint, and micro-movement in real time. 25 degrees of freedom. Complete, accurate, low-latency data for high-quality datasets and advanced applications.

Precision data gloves that bring human dexterity to robotics and motion capture.

MANUS data gloves capture every finger, joint, and micro-movement in real time. 25 degrees of freedom. Complete, accurate, low-latency data for high-quality datasets and advanced applications.

How does human hand motion become robot training data? Across @NVIDIARobotics's GR00T N1, EgoScale, and GR00T 2 research, MANUS gloves provide precise hand-motion data to support robot teleoperation, human–robot alignment, and learning from human demonstrations. Our latest blog explores how these workflows support Vision-Language-Action (VLA) and World Action Model (WAM) training, and what robotics teams should consider when building their own data collection pipeline. Read more: manus-meta.com/blog/robot-da… #MANUS #RobotLearning #PhysicalAI #Robotics #NVIDIA

Congratulations to Genesis AI @gs_ai_ on the $500M Series A! Great to see MANUS data gloves in action in the announcement video.🤖
Genesis is entering its next phase. Today we've raised a $500M Series A at $3.5B. We can imagine far more than we can build physically today. Almost everything we are creating is still paid for in human time, by human hands. We want to remove the physical limits on human ambition. We are building general-purpose robots with human level intelligence and capability. Ones that work reliably across entire workflows for years. There are countless versions of possible futures where billions of robots are deployed everywhere, and we aspire to bring the brightest one that carries our hopes into reality.

NVIDIA GR00T 2 introduces a cross-adaptation capability, X-Adapt, that enables robots to learn new dexterous skills without task-specific teleoperated demonstrations. Operators perform tasks naturally while MANUS data gloves capture precise finger-joint motion, synchronized with wrist tracking and multi-view video. The recorded motion is then retargeted to the robot’s embodiment, preserving key fingertip relationships for grasping and pinching. Combined with task-agnostic robot data, these demonstrations help GR00T 2 learn new skills across different embodiments. By making natural human demonstrations more accessible for robot training, X-Adapt offers a promising path toward scaling dexterous robot learning. Explore our latest blog to learn how MANUS contributes to this process, from human motion capture to robot learning. Read the full article: manus-meta.com/blog/nvidia-g… #GR00T2 #RobotLearning #NVIDIA #PhysicalAI #Robotics #MANUS

That’s a wrap on #WebSummit Lisbon! 👋 A big shoutout to everyone who stopped by our booth to see Flex 2 in action, and catch Prima 1’s European debut. We loved seeing so many of you have fun during the live demos! It was great meeting you all at Web Summit this year. Obrigado, Lisboa!

In a recent study from @ADI_News, researchers demonstrated how tactile sensing can enable robots to perform a complex manipulation task without relying on vision. In "Touch2Trace", a robotic hand was trained to feed cables through its fingers using tactile feedback and joint-state data, achieving zero-shot transfer to unseen cables and routing conditions. To collect training demonstrations, operators used @ManusMeta Metagloves Pro to capture finger movements and teleoperate a @Tesollo_Inc DG-5F robotic hand. Joint-state and tactile data were recorded throughout the process for training. Read the paper: arxiv.org/pdf/2609.15921 #humanoids #humanoidrobotics #humanoidrobot #humanoid

During @DrJimFan’s keynote at #CoRL2026, @NVIDIARobotics showcased #GR00T 2 and how human demonstrations can teach robots new dexterous skills. @ManusMeta data gloves are part of NVIDIA’s human demonstration setup, capturing finger-joint motion alongside wrist tracking and head- and wrist-mounted cameras. NVIDIA’s pipeline then maps the recorded motion to robot arms and hands for training. From human hand motion to robot learning, we’re proud to see MANUS contribute to this next chapter of Physical AI. Learn more: research.nvidia.com/labs/gea… #MANUS #NVIDIA #GR00T #PhysicalAI #RobotLearning

MANUS Core 3.2 is here. This release brings Linux ARM64 support, more accurate and consistent pinch tracking across a wide range of hand sizes, a new Python SDK, and haptic feedback support in NVIDIA Isaac Teleop. Download MANUS Core 3.2: docs.manus-meta.com/3.2.1/Re… Read the release notes: docs.manus-meta.com/3.2.0/Re…

How do robots learn human dexterity? 🤖 With precise hand data. We're building the technology behind robot dexterity. Our latest Core 3.2 update makes it easier for developers to integrate our data gloves into robotics workflows. Follow @ManusMeta for more Physical AI demos.

One less PC between your hands and your robot. MANUS Core 3.2 brings Linux ARM64 support, making it easier to integrate hand tracking directly into your robotics setup for teleoperation and data collection. See what this unlocks in 60 seconds. #MANUS #Robotics #Teleoperation #EmbodiedAI

🎃Things are getting out of hand. Luckily, we know a Thing or two about hand tracking. 🖐️ @AltBionics @RaptorHouseFX #Halloween #MANUS

We're proud to announce the winners of the Humanoid Robotics Industry Awards 2026! Your winners: ⭐ Industry Leadership: @theermann, @dynamicreatures, Founder & CEO ⭐ Outstanding Company: @TheHumanoidAI ⭐ Outstanding Research Organisation: @BoschGlobal Center for Artificial Intelligence ⭐ Groundbreaking Technology: @ManusMeta ⭐ Best Use of AI: @AGIBOTofficial From everyone at Humanoid Robotics Technology and our awards partner Novanta Inc.: a big thank you to everyone who nominated and voted this year, and a massive congratulations to all our winners and finalists for their outstanding contributions to advancing humanoid robotics and shaping the future of the industry. Check out the full announcement here for more info: humanoidroboticstechnology.c… #humanoid #humanoidrobots #humanoidrobot #humanoids

Needs more cowbell
Tesla’s Optimus training regime leak. Workers in its Palo Alto R&D facility have been holding “robot university” for many months already. Press got it wrong when they said it just started up. Workers are wearing 24 lb suits with up to 40 cameras and grids of 3D sensors. To capture very detailed video and 3D fields of whatever the worker does. There are multiple configurations. They are given different tasks to do. They don’t know where the tasks come from. Multiple squads are working on three shifts 24 hours a day. Similar efforts were seen at the Texas factory. The data collection effort is immense. A Tesla car only has nine cameras. This leads to a data flywheel that is orders of magnitude more data than the car captured. Separate teams are capturing home data, like making a sandwich, while other teams capture jobs on factory floor, like installing HVAC systems in cars. Elon isn’t playing around. I can’t tell you how I learned this.

🏆 WE WON! 🏆 MANUS is the winner of the Groundbreaking Technology Award at the 2026 Humanoid Robotics Industry Awards! After 12 years of developing technology that translates human hand movements into precise and reliable data, we’re super proud to see our work recognized by the robotics community. As our Co-founder & CEO @stephan__brink puts it: “Building something unique and valued is very hard, so receiving this recognition from the industry is thoroughly heartwarming. We are nowhere near done with our vision, and the coming year will be especially exciting.” A huge thank you to our incredible team, our customers and partners, and everyone who voted for us. ❤️🤖 @HumanoidRTech

"What I really like about MANUS is that we make a product from beginning to end. We design it, we build it, we test it, we produce it, we ship it." At @ManusMeta, you don't just work on one piece of the puzzle. You see your work go from first sketch to a customer's hands. We're building the technology that teaches robots human dexterity, and we're hiring. Check out our open roles: manus-meta.com/careers

We've been running a fully autonomous demo of FLUX-mimic nonstop at GTC for two days. We're even golden hour -proof. Lfg @mimicrobotics