Ecosystem

Unified Stack for Sim & Real Teleoperation

A single framework that works seamlessly across simulated and real-world robots, ensuring streamlined device workflow and consistent data schemas.

Robotics Stacks

Stack	Description	Status
ROS2	Widely adopted middleware for robot software integration and communication	Supported
Isaac Sim	Simulation platform to develop, test, and train AI-powered robots	Supported (v6.0)
Isaac Lab	Unified framework for robot learning designed to help train robot policies	Supported (v3.0)
Isaac ROS	NVIDIA CUDA-accelerated toolkit for ROS2	Planned
Isaac Arena	Isaac Lab extension for large-scale evaluation and resource orchestration	Planned (v0.2)

Supported Input Devices

Isaac Teleop provides a standardized interface for teleoperation devices, removing the need for custom device integrations and ongoing maintenance. It supports multiple XR headsets and tracking peripherals. Each device provides different input modes, which determine which retargeters and control schemes are available. Easily extend support for additional devices through a plugin system; see Isaac Teleop device plugin (C++ level) for details.

XR Headsets and Tracking Peripherals

Device	Input Modes	Client / Connection	Notes
Apple Vision Pro	Hand tracking (26 joints), spatial controllers	Isaac XR Teleop Sample Client (visionOS app)	Build from source; see Connect Apple Vision Pro
Meta Quest 2/3/3S	Motion controllers (triggers, thumbsticks, squeeze), hand tracking	Isaac Teleop Web Client (browser)	See Connect Quest and Pico
Pico 4 Ultra	Motion controllers, hand tracking	Isaac Teleop Web Client (browser)	Requires Pico OS 15.4.4U or newer
Pico Motion Tracker	Full body tracking	Isaac Teleop Web Client (browser)	Requires Pico OS 15.4.4U or newer Requires Pico Browser 4.0.40 or newer (Enterprise enabled)

In addition to the fully integrated XR headsets, Isaac Teleop also supports standalone input devices. Those devices are typically directly connected to the workstation where the Isaac Teleop session is running via USB or Bluetooth. See Isaac Teleop device plugin (C++ level) for more details.

Standalone Input Devices

Device	Input Modes	Client / Connection
Manus Gloves	High-fidelity finger tracking (Manus SDK)	Manus Gloves Plugin (CLI tool)
Logitech Rudder Pedals	3-axis foot pedal	Generic 3-axis Pedal Plugin (CLI tool)
OAK-D Camera	Offline data recording	OAK-D Camera Plugin (CLI tool)
Haptikos Exoskeletons	High-fidelity finger tracking and haptic feedback (Haptikos Robotics API)	Haptikos Plugin (CLI tool)

Planned Input Device Support

The following input devices and device categories are planned for support in the future:

Planned Input Devices

Device	Input Modes	Client / Connection	Status
JoyLo	Master Manipulators	CLI tool with USB connection	Planning, see #272
Gello	Master Manipulators	CLI tool with USB connection	Planning, see #273
Haply	Master Manipulators	CLI tool with USB connection	Planning, see #274
SO-101	Master Manipulators	CLI tool with USB connection	Planning, see #275
3D Space Mouse	HID input	CLI tool with USB connection	Planning, see #276
Gamepad	HID input	CLI tool with USB/Bluetooth connection	Planning, see #277
Keyboard	HID input	CLI tool with USB/Bluetooth connection	Planning, see #278

Targeted Robotics Embodiments

• Retarget the standardized device outputs to different embodiments.

• Reference retargeter implementations, including popular embodiments such as Unitree G1.

• Retargeter tuning UI to facilitate live retargeter tuning.

Device Acquisition

For inquiries about acquiring supported or planned devices, please contact the manufacturers directly. Each device name in the tables above links to the corresponding manufacturer page.

Manufacturer Contacts

Manufacturer	Devices	Acquisition Contact
Apple	Apple Vision Pro
Meta	Meta Quest 2/3/3S
Pico (ByteDance)	Pico 4 Ultra, Pico Motion Tracker	Veronica Li Email: Veronica.li@bytedance.com Mobile: +1 (909) 569-2774
Manus	Manus Gloves
Logitech	Logitech Rudder Pedals
Luxonis	OAK-D Camera
Haptikos	Haptikos Exoskeletons
Haply	Haply (planned)
3Dconnexion	3D Space Mouse (planned)

Social Links

Join the Isaac Teleop community to ask questions, share your work, and stay up to date.

Discord (NVIDIA official)

Isaac Teleop has a dedicated isaac-teleop channel in the official NVIDIA Omniverse Discord server. Joining takes two steps:

1. Join the server. Open the NVIDIA Omniverse Discord invite and accept it to become a member of the NVIDIA Omniverse Discord server.

2. Open the channel. Once you are a member of the server, go to the #isaac-teleop channel to reach the team and the community. (The channel link only works after you have joined the server in step 1.)

WeChat (微信) group (community-run)

A community-driven WeChat group is also available. Scan the QR code below from within the WeChat app (Discover → Scan) to join. Scanning with your phone’s built-in camera app will not work.

Figure: Scan from within the WeChat app (not the camera app) to join the community WeChat group

Note

This WeChat group is created and maintained by community members. It is not an official NVIDIA channel. The NVIDIA team does not participate in group operations, but does monitor the group and answer questions when possible.

More than one community chat group may exist, and we may list additional groups here as we become aware of them. That said, NVIDIA encourages the community to organically converge on a single chat group, so that discussions stay consolidated in one place.