RoboParty has launched RP1, an open-source humanoid robot platform for robotics research and embodied AI development. Researchers can work across its hardware and software, rather than being limited to a closed interface, giving them a way to investigate where a robot’s behaviour comes from and how to change it.
Watch Desk analysis
What happened
The Shanghai-based company offers RP1 in two configurations: a 1.3-metre Standard model with 24 actuated degrees of freedom, and a Pro version with 30. The AI Insider reports that the platform includes modular hardware, ROMOMO actuators, control software, firmware, simulation tools and robot-learning frameworks. The company says RP1 is available through GitHub.
The toolkit also supports VR teleoperation, motion capture, data collection and playback, motion retargeting, and whole-body control, which coordinates the robot’s arms, torso and legs. RoboParty says researchers can take work from data collection and simulation training through to deployment on a physical robot, then compare model outputs with what the machine actually does. That could help teams trace a failed movement to the model, control policy, actuator or mechanical design, rather than guess at the culprit from the outside.
Why it matters
Robotics research often has to bridge simulation and physical machines, where hardware differences can make experiments harder to reproduce. RoboParty says it has focused on consistency in the robot’s structure, actuators and manufacturing so researchers can compare results more reliably. An open platform with tools across that stack could make it easier for teams to inspect and modify the whole system, not just its code.
The release builds on RoboParty’s RPO open-source project, launched in January. The company says that project has received more than 2,600 GitHub stars. That shows interest in the earlier project, not proof that RP1 has already become a research standard.
Our read
This is a meaningful robotics launch because the openness extends beyond a software repository: researchers are being offered a route into the robot’s hardware, controls and simulation workflow too. The useful test will be whether other teams can reproduce experiments and make changes without the platform’s promised flexibility turning into a hardware-and-software scavenger hunt.
What to watch
- Which hardware designs, software and learning resources are actually published and usable.
- Whether independent research teams can reproduce results across RP1 units.
- How well work developed in simulation transfers to the physical robot.
Discussion spark: For robotics research, is an open hardware-and-software stack more valuable than a more capable but closed robot platform?
Sources and evidence
- China’s RoboParty Launches Open-Source RP1 Humanoid for Robotics Research (15 October 2026, 21:15 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.