OpenPyRo-A1: An Open Python-Based Low-Cost Bimanual Robot for Embodied AI
Helong Huang, Christopher Edwin Mower, Guowei Huang, Sarthak Das, Magnus Dierking, guangyuan Luo, Kai Tan, xi Chen, Yehai Yang, Yingbing Chen, Yiming Zeng, Yinchuan Li, zhanpeng ZHANG, Shuang Wu, Yingxue Zhang, Weichao Qiu, Tongtong Cao, Mian Qin, Sajjad Pakdamansavoji, Yuecheng Liu, Yuzheng Zhuang, Guangjian Tian, Jianye Hao, Jun Wang, Haitham Bou Ammar, Xingyue Quan
AI summary
Problem
Bimanual robotics research is bottlenecked by the high cost of dual-arm platforms and a severe scarcity of high-quality manipulation datasets, which limits the development of scalable embodied AI.
Approach
The authors engineered a modular, ~$14K bimanual robot and a lightweight Python-based distributed control system that integrates VR teleoperation, real-time safety constraints, and seamless data logging for policy training.
Key results
- Sub-0.2 mm positional repeatability with 5 kg payload per arm
- Low-latency (<10 ms) Python-first control framework with built-in safety constraints
- Curated dataset of over 560 bimanual manipulation episodes via VR teleoperation
- Successful deployment of imitation learning and agentic VLA/LLM policies on real hardware
Why it matters
It democratizes dual-arm robotics research by providing an accessible, high-performance platform for scalable data collection and embodied AI development.
Abstract
Many real-world tasks, such as assembly, cooking, and object handovers, require bi-manual coordination. Learning such skills via imitation remains challenging due to dataset scarcity, mainly caused by the high cost of bi-manual robotic platforms and barriers to entry in robotics software. To ad- dress these challenges, we introduce (1) OpenPyRo-A1, a low- cost, bi-manual humanoid robot priced at approximately $14K. OpenPyRo-A1 achieves 0.2 mm repeatability and supports a 5 kg payload per arm, and (2) a Python-first distributed control framework for seamless teleoperation, data collection, and policy deployment, designed for ease of use; moreover, the code-base is installable via pip. We conducted imitation learning experiments in both simulation and the real world, integrating the robot with perception models, motion planning, and a large language model. The results demonstrate that OpenPyRo-A1 is a stable, user- friendly, and high-precision dual-arm platform. We expect that the OpenPyRo-A1 hardware, control system, and curated dataset of bi-manual manipulation episodes will advance affordable and scalable dual-arm robotics.