AirExo profile: a low-cost exoskeleton for whole-arm teleoperation
A wearable leader arm at about $300 per side that can also record demonstrations with no robot in the room.
Contents
| Type | Open-source arm exoskeleton |
|---|---|
| Founded | Published at ICRA 2024 |
| Pricing | Published |
| Website | airexo.github.io |
Desktop leader arms tie the operator to a bench. AirExo, from Shanghai Jiao Tong University and the Shanghai Artificial Intelligence Laboratory, puts the leader on the operator’s arms instead1.
Best for: Whole-arm demonstrations on a research arm, including collection away from the robot.
Not the right fit for: Teams that want a commercial product, or finger-level capture.
How it works
AirExo is a portable dual-arm exoskeleton with a one-to-one joint mapping to the robot, which gives joint-level teleoperation. The authors put the cost at approximately $300 per arm. It was developed for Flexiv Rizon arms and, they state, can be quickly modified for others including the UR5, Franka and Kuka. Worn away from the robot, it collects demonstrations in the wild1.
Published evidence
In the paper, presented at ICRA 2024, a policy trained on 10 teleoperated demonstrations, about 3 minutes, plus in-the-wild data surpassed one trained on 50 teleoperated demonstrations, over 20 minutes, on a grasping task1. The result covers the authors’ tasks. Related designs exist for other robots: HOMIE pairs exoskeleton arms with gloves and a pedal for humanoids at a stated $5002, and ACE combines an exoskeleton with a hand-facing camera3.
Strengths and limits
Strengths
- Joint-level control like a leader arm, but worn, so the operator’s whole arm does the work.
- Can record demonstrations with no robot present, which removes the robot as the bottleneck.
- The authors report that 3 minutes of teleoperated data plus in-the-wild data matched or beat more than 20 minutes of teleoperated data on their tasks.
Limits
- A research build: no vendor or support.
- Developed for one arm family; other arms need modification.
- The reported results are from the authors’ two tasks, not a general benchmark.
Choose another option if: A desktop leader arm is enough (GELLO), or you are teleoperating a humanoid (HOMIE’s exoskeleton cockpit or a headset).
Pricing
Approximately $300 per arm, as reported by the authors.
Questions to settle before building
- Is our arm one the design already fits, or will we modify it?
- Will in-the-wild collection match the robot’s workspace and camera view?
- How is the exoskeleton fitted to different operators?
- Who repairs it when a joint fails mid-campaign?
See how AirExo compares with the other options in the rig comparison table.
Sources
- AirExo: low-cost exoskeletons for learning whole-arm manipulation in the wild. Fang et al. (ICRA 2024). Accessed 11 Nov 2026.
- HOMIE: humanoid loco-manipulation with isomorphic exoskeleton cockpit. Ben et al. (RSS 2025). Accessed 11 Nov 2026.
- ACE: a cross-platform visual-exoskeleton system for low-cost dexterous teleoperation. Yang et al. (CoRL 2024). Accessed 11 Nov 2026.