Browse our latest products
Frequently asked questions
1) What’s the difference between ROVA 5 and ROVA 6?
Answer:
Both are SO-101–based arms built for learning, teleoperation, and AI data collection. The key difference is dexterity.
- ROVA 5 is simpler to control, faster to learn, and great for pick-and-place, sorting, basic manipulation, and “first robotics arm” use cases.
- ROVA 6 adds an extra joint for better orientation control (more human-like movement), making tasks like angled grasps, reaching around objects, and more precise pose control easier.
In both versions, the “+1” refers to the gripper.
2) Can I use ROVA 5 and ROVA 6 with ROS / ROS2?
Answer:
Yes. You can integrate the arm into ROS or ROS2 workflows for motion control, camera streams, and higher-level autonomy. Common ROS patterns include:
- Publishing joint states and commanding joint targets
- Integrating a wrist/overhead camera pipeline
- Using motion planning frameworks (where applicable) and writing your own controllers for smooth motion
If you’ve worked with ROS topics, services, and URDFs before, you’ll feel at home.
3) Do ROVA 5 & ROVA 6 arms work with the LeRobot library (Hugging Face)? What can I do with it?
Answer:
Yes. These arms are designed to be compatible with LeRobot-style workflows:
- Teleoperate the arm (leader → follower)
- Record demonstrations (actions + timestamps + camera frames)
- Build clean datasets for learning-based control
- Train / fine-tune policies and run closed-loop inference for real robot behavior
This is the most practical path to go from “manual teleop” → “robot does it on its own.”
4) What software skills do I need to get started with ROVA 5 & ROVA 6?
Answer:
You can start at different levels:
- Beginner: run sample scripts, move joints, open/close gripper, stream camera, and do teleop recording.
- Intermediate: tune motion smoothing, add calibration routines, and build dataset pipelines for training.
- Advanced: ROS2 integration, custom control loops, imitation learning / RL, and VLA-style experiments.
If you can do basic Python and follow setup instructions, you can start moving the robot the same day.
5) What kind of tasks are ROVA 5 & ROVA 6 Arms best suited for?
Answer:
These arms are optimized for tabletop manipulation + AI training. Typical demos people build quickly:
- Pick-and-place, sorting, stacking, simple assembly
- “Move to a visual target” using the wrist/overhead camera
- Data collection for imitation learning (teleop → dataset → trained policy)
They’re ideal for enthusiasts, labs, and students who want hands-on experience with real robot data, not just simulation.