The eyes of Embodied AI. Spatial AI cameras with onboard VIO/VSLAM, robot navigation & embodied AI data capture. Built for robotics developers.

Wan Chai, HK
Joined December 2025
Get a sneak peek at Insight 7 in our official teaser video. 🎬 Designed for compact mobile robotic platforms, Insight 7 features a 70 mm standard stereo baseline and an ideal range of 0.15 m to 15 m, optimized for mid-range perception. With a more compact form factor, Insight 7 inherits the same onboard VIO, 10 TOPS AI acceleration engine, and ROS integration capabilities as Insight 9. Equipped with Dual Global Shutter sensors, it delivers synchronized visual, spatial, and motion data out of the box. Coming soon. 🦾 #SpatialAI #Robotics #ComputerVision #EmbodiedAI #ROS #VSLAM
1
4
675
See ahead. Keep more ground in view. 🤖 With a 157° vertical and 188° diagonal RGB field of view, Insight 9 captures the ceiling, the sofa ahead, and the nearby floor toy in a single frame. This RGB comparison shows what broader vertical coverage brings to robot vision: more of the surroundings in frame and more flexibility in camera placement. #Robotics #ComputerVision
60
2 minutes of rapid robot arm circles. The position endpoint error? Just 7.18 mm. 🤯 Watch Insight 9 move through repeated circles and return to the start, with the physical motion and VIO-estimated trajectory shown side by side. Insight 9 delivers pose data directly from the camera, combining global-shutter stereo vision, an IMU and onboard VIO with factory calibration and time synchronization. Less localization front-end integration work gives your team more time to develop the robot application. #Robotics #VIO
1
68
All-in-One Smart Camera. From perception to robot navigation. This Lite robot demo brings perception and vision-only navigation together inside Insight 9. In this demo configuration, Insight 9 provides stereo depth and onboard VIO/VSLAM pose data. TinyNav runs on the device for mapping, map-based localization, path planning, and obstacle avoidance. Select a goal on the phone, then watch the robot navigate around obstacles and continue toward its destination. An ESP32 and motor driver board handle communication and chassis motion control. Perception and navigation compute stay on Insight 9, with no separate navigation computer used in this demo. #Robotics #ComputerVision
86
💻 ROS 2 on macOS, with Insight 7. From a prepared ROS 2 environment to live RGB and stereo camera views, this quick start brings Insight 7 into your Mac development workflow. Stereo depth and visual-inertial localization run onboard, reducing host-side processing and integration work. Demonstrated on a Mac mini. A configured ROS 2 environment is required. #ROS2 #Robotics
3
7
605
Cup stacking calls for repeated grasping, alignment and placement. Clothes folding requires bimanual adjustment as fabric changes shape. In embodied AI data collection, hand motion is more informative when linked to object state and task progression. #EmbodiedAI #Robotics
1
105
Three scenes that expose depth map weaknesses: a large low texture surface under strong light, layered foliage, and reflective leaves with complex edges. Captured side by side, Insight 9 shows more continuous depth and fewer visible gaps across the tested regions than the comparison camera. Insight 9 also integrates RGB, IMU and onboard VIO/VSLAM for robot perception and localization. #Robotics #ComputerVision
1
2
26
1,447
⚡️ Ubuntu + ROS 2. Meet Insight 7. Get started with live RGB and stereo camera views in a prepared ROS 2 environment. Onboard stereo depth and visual-inertial localization help reduce host-side processing. Watch the Ubuntu quick-start guide. #ROS2 #Robotics
2
7
30
3,908
Windows 11 desktop. Linux ROS 2 toolchain. Live Insight 7 data in one workflow. This quick start moves from unboxing and ROS mode selection to live RGB and stereo image streams in rqt through WSL 2 and Ubuntu 22.04. With stereo depth and VIO/VSLAM running onboard, Insight 7 helps reduce host-side integration for front-end perception and localization. Use the Bring-Up Guide that matches your firmware. #ROS2 #Robotics
153
That’s a wrap on #ROSCon2026 in Toronto! 🙌 Thank you to everyone who stopped by our booth at Startup Alley, SU-3, joined our stage showcase, and shared your projects, questions and ideas with us. It was great to connect with the @OpenRoboticsOrg community and be here alongside our partner @DRoboticsDev. We’re heading home with new ideas, valuable conversations and plenty of inspiration for what comes next. See you at the next one! 🤖 #ROS
1
1
7
577
Good morning from the final day of #ROSCon2026 in Toronto! 🤖 We’re back at Startup Alley, SU-3, for more conversations with the @OpenRoboticsOrg community. Stop by and say hello. #ROS
3
7
518
A first look at our booth at #ROSCon2026! 🤖 LooperRobotics is in Toronto with Insight 9, Insight 7 and Insight 3 Spatial AI Cameras, EgoHalo H1 for embodied AI data collection, and NavCore for robot navigation. Great to be here with our partner @DRoboticsDev and connect with the @OpenRoboticsOrg community! 📍 Startup Alley · SU-3 🎤 Robot demo: Sept 23, 12:45–13:00 EDT Stop by, meet the team and let’s talk robotics. #ROSCon #ROS #Robotics
1
1
4
422
LooperRobotics retweeted
🔥ROSCon Global 2026 — we made it, did YOU? 👀 Catch @DRoboticsDev at our partners’ booths: 📍 @AgilexRobotics Booth #48 📍 @LooperRobotics Booth SU-3 | Startup Alley RDK, #ROS 2, #Robotics, #EmbodiedAI...— come talk to us, we're so ready 🤙 #ROSCon2026
1
1
4
201
LooperRobotics retweeted
We’re thrilled to announce that D-Robotics is a sponsor of ROSCon Global 2026. From ROS 2 applications to intelligent machines in the real world — D-Robotics is bringing RDK to @OpenRoboticsOrg #ROSCon Global 2026! For us, ROS 2 is not an add-on — it is a core part of the RDK robotics development experience. Across the RDK platform, we support multiple ROS 2 distributions, while TogetheROS.Bot builds on ROS 2 with hardware-accelerated capabilities tailored for robotics. At ROSCon, we’ll be showing how RDK dev kits + ROS 2 can help developers bring edge AI from the development environment to real robotic systems, with RDK providing the computing foundation to run AI workloads closer to the robot. Find D-Robotics at our partners' booths ⬇️: 📍 AgileX Robotics Booth #48 📍 LooperRobotics Booth SU-3|Startup Alley And you can't miss our Lightning Talk by D-Robotics Engineer Shukun Huang 🎤 AI-Native + ROS: A New Paradigm for Robot Development Got a ROS 2 project in the works? Bring your ideas, your questions, or even your toughest robotics problem. Come find us, say hi, and let’s build something interesting together. See you in Toronto! en.d-robotics.cc/blog/d-robo…
4
9
725
#ROSCon2026 is almost here! LooperRobotics is bringing Insight 9, Insight 7 and Insight 3 Spatial AI Cameras, EgoHalo H1 for embodied AI data capture, and NavCore for robot navigation to Toronto. Meet us and connect with the @OpenRoboticsOrg community at Startup Alley, SU-3. Our partner @DRoboticsDev will be there too. Looking forward to catching up in Toronto! 📍 Toronto · Sept 22–24 📌 Startup Alley · SU-3 🎤 Stage showcase · Sept 23 #ROSCon
2
4
538
One task, three complementary views. Looper Data Acquisition Package combines a head-mounted EGO view and dual wrist-mounted UMI views with RGB-D, IMU, and VIO camera pose data. See calibration, hand-joint visualization, and manipulation demos. #Robotics #EmbodiedAI
1
2
499
Two Insight 9 cameras run previous and updated versions side by side. This comparison shows reduced depth-map flicker after the update. Insight 9 uses learning-based stereo depth estimation for robotic perception. #ComputerVision #Robotics
2
8
964
Heading to #ROSCon2026? Let’s meet! LooperRobotics is excited to connect with the @OpenRoboticsOrg community to discuss VIO/VSLAM, sensor sync, embodied AI data collection, ROS integration, and what you’re building. Toronto · Sept 22 to 24 Startup Alley · SU-3 #ROSCon
2
2
8
721
During repeated tabletop impacts, the peak absolute value on a single acceleration axis reached approximately 130 m/s², exceeding 12 g. Insight 7 is a compact spatial perception camera with a 6-axis high-dynamic IMU. Its ±24 g accelerometer measurement range was selected to meet the demands of high-dynamic motion while maintaining IMU measurement accuracy. #Robotics #IMU
1
1
174
Before NavCore: camera, compute, router, power, cables, and wired debugging. After: NavCore and a phone. Wi-Fi, two RJ45 ports for two LiDAR sensors, 9-36 V input, 12 V output, Insight 9, NVIDIA Jetson, and TinyNav bring perception, mapping, and navigation together. #RobotNavigation #UnitreeGo2
3
11
492
Fast circular motion. Continuous onboard VIO. Motion blur, inter-frame displacement, and rapidly changing viewpoints put visual-inertial tracking under pressure. Rotation after rotation, Insight 7’s onboard VIO output traces a dense, connected multi-loop 3D path with no visible break or tracking reset. The computer only receives and visualizes the result. #VSLAM
2
157