Unified Physical AI framework · ROBOTIS
Cyclo is a modular framework for building and operating Physical AI systems. It follows ROBOTIS’s open-source philosophy: put the building blocks in the open so Physical AI is easier for anyone to learn, integrate, and run on real hardware—not a closed, single-vendor path. It covers AI integration, action-data workflows, robot control, simulation, and operations. Private infrastructure (supervision, proprietary data, and similar) can plug in alongside the open public stack.
Physical AI Lineup
Actuator products
- DYNAMIXEL series: DXL, DXL-X, DXL-P, DXL-Y, DXL-Q — interface stack: Dynamixel SDK
Public modules, optional private stack, Physical AI lineup, and DYNAMIXEL actuator products.
| Module | Repository | Role | Release |
|---|---|---|---|
| Cyclo Manager | cyclo_manager |
Operations and system management |  |
| Cyclo Intelligence | cyclo_intelligence |
Next-generation imitation learning workflows |  |
| Cyclo Control | cyclo_control |
Whole-body control and execution |  |
| Cyclo Lab | cyclo_lab |
Simulation and Reinforcement Learning |  |
| — | robotis_interfaces |
Shared interface definitions |  |
| — | robotis_applications |
Application-level integrations and third-party dependencies |  |
Additional private or product-specific repositories may exist depending on deployment.
Not all components are part of the public release. Typical internal extensions:
- Cyclo Supervisor — robot supervision
- Cyclo Hub — private data infrastructure
| Resource | Description | Link |
|---|---|---|
| Documentation | Product and software manuals | Manual |
| Discord | Community chat and support | ROBOTIS Community |
| YouTube | Tutorials and demos | ROBOTIS Open Source Team |
Contributions are welcome via issues and pull requests in the relevant Cyclo repositories.