Toborlife AI Launches G1 Teleoperation Stack

Recap

Source: 로봇신문, report of Oct. 2, 2026

According to the report, Toborlife AI, the developer of the Tobor Harness control system and a US partner of Unitree Robotics, has launched a full-body teleoperation software package for the Unitree G1 Edu humanoid. This release provides US-based research institutions, universities, and developers with what the company describes as the first data collection, learning, and support stack for the G1 that is built and supported domestically. A key feature of the system is its data sovereignty policy: all learning data collected through the platform remains exclusively on the customer's hardware and is never transmitted to external servers or under the control of Unitree.

The system utilizes PICO virtual reality headsets to allow operators to control the robot's arms, legs, and torso in real-time without writing intermediate scripts. When equipped with the G1's optional dexterous hands, operators can perform finger-level control for precise grasping and fine motor tasks. The software includes built-in data collection that records full-body movements in a format ready for immediate model training, rather than raw logs requiring extensive processing. Toborlife AI supports these deployments with a US-based team of robotics engineers who provide direct technical support and on-site training.

Geoffrey Eakas, CTO of Toborlife AI, stated that the system is designed to help teams focus on specific applications by providing a robust foundation for data collection. He noted that the company continuously optimizes the tools, allowing research teams to improve their models over time without additional manual effort. David Schulhof, CEO, emphasized that the company's mission is to bridge the gap between humans and robots, making the technology accessible to a wider audience regardless of their technical background.

Context

The Unitree G1 has become a popular choice for research labs due to its cost-effectiveness and capabilities, but it has often been criticized for lacking a mature software ecosystem in the West. Researchers frequently spend months integrating various open-source libraries to get the robot to perform basic tasks. This launch addresses that gap by providing a turnkey solution for teleoperation and data collection, which is a critical step in training reinforcement learning models for humanoid robots.

Robot's take

This launch is significant because it lowers the barrier to entry for US researchers who want to work with Chinese-made hardware without compromising data security. By ensuring data stays local, Toborlife AI addresses a major concern for institutions handling sensitive research. However, the long-term success of this platform will depend on how well it integrates with other popular robotics frameworks and whether the teleoperation interface remains intuitive for non-experts. It is not yet clear if this software will be the standard for G1 development or if other competitors will emerge with similar offerings.