Robots today are physically capable but cognitively limited. They can move, but they can’t act intelligently. To make them useful, they need perception, planning, and control — skills traditionally built through complex, local pipelines.
GRID changes that.
GRID is a cloud-native platform that provides AI capabilities to robots through simple APIs. It removes the infrastructure burden of deploying perception and planning models, enabling developers to control and automate robots directly from the cloud.
No setup. No local GPU. No ROS debugging.
AI Skills for Trossen Robots on GRID!
With just a few lines of code, all Trossen arms, whether existing or newly purchased, can now access AI skills. We have integrated the following form factors from Trossen Robotics:
- Standalone Arm: Trossen WidowX AI
- Bimanual Mobile: Trossen Mobile AI
- Bimanual Fixed: Trossen Stationary AI
Teleoperated Control
GRID enables teleoperation for scalable data collection.
Try it out on on NVIDIA Isaac Sim through this Open GRID session and deploy easily on real hardware.
Playback on any Trossen arm:
driver_left = trossen_arm.TrossenArmDriver()
driver_right = trossen_arm.TrossenArmDriver()
driver_left.set_arm_positions(left.values(), dt, False)
driver_right.set_arm_positions(right.values(), dt, False)Record data in Sim:
from grid.robot.arm.isaac_bimanual import IsaacBimanual
agent = IsaacBimanual()
left_arm_pos = agent.larm.getJointAngles()
right_arm_pos = agent.rarm.getJointAngles() Pick and Place via API
GRID provides plug and play skills like Grasp Detection and Motion Planning. Stay tuned for a demo and sample GRID notebook!
These can be accessed via simple code:
from grid.robot.arm.widowx import WidowX
arm = RobotArm(serial_number="YOUR_TROSSEN_SERIAL")
grasp = arm.skills.detect_grasp(image=camera_frame)
trajectory = arm.skills.plan_pick_and_place(grasp=grasp)
arm.execute(trajectory)Don’t own a Trossen Hardware?
Try it on NVIDIA Isaac Sim through Open GRID: grid.generalrobotics.dev
Want to get your robot on GRID?
We partner with manufacturers to ingest 3D models and system specifications to build a simulation-ready digital twin, allowing us to test and validate AI skills before real-world deployment. This dramatically accelerates integration and unlocks AI capabilities—like teleoperation and pick-and-place—across diverse hardware platforms like Trossen arms.
Contact General Robotics for more info: info@generalrobotics.company
Bridge the gap between robots and intelligence with GRID making robots useful — fast!
.webp){kind=small}
