Apple Vision Pro

Apple's head-mounted display, whose hand tracking embodied-AI researchers commonly use for teleoperation and collecting human data.

Apple Vision Pro is a headset Apple unveiled at WWDC in June 2023 and began selling in the US on February 2, 2024, starting at $3,499; it originally shipped with an M2 chip plus an R1 chip dedicated to processing sensor input, and has 12 cameras, 5 sensors, and 6 microphones, with an M5 version following in October 2025. It is not a robotics product, but its system tracks the wrist and fingers' 3D pose in real time and has a high-resolution stereo display, so embodied-AI research has adopted it for teleoperation: a human's hand motion, after retargeting, drives the robot, and the robot's head camera streams back to the headset. It can also record egocentric video and 3D hand trajectories in sync while a person does chores, which is used to collect human manipulation data.

ExampleOpen-TeleVision used a Vision Pro to remotely control a Unitree H1 — the author, at MIT, controlled a robot physically located at UC San Diego; Apple itself used Vision Pro to record EgoDex, an 829-hour, 194-task egocentric dataset of human hand manipulation.

Also called

AVP

Related

VR Teleoperation · VR Headset · Open-TeleVision · EgoDex · Egocentric Video · Meta Quest 3 / PICO 4 Ultra XR Headsets

Sources

Apple Newsroom: Apple Vision Pro available in the U.S. on February 2
Wikipedia: Apple Vision Pro
Hoque et al. 2025: EgoDex (arXiv 2505.11709)

As of

2025-10