OpenArm 101: The Open-Source Robot Arm for Research (2026 Guide)

What Is OpenArm 101?

OpenArm 101 is a 6-degree-of-freedom open-source robot arm developed and sold by Robotics Center of Silicon Valley (SVRC). It was designed from the ground up for two specific applications: teleoperation data collection and imitation learning research. Most research robot arms are derived from industrial platforms, which optimizes for repeatability and payload at the expense of back-drivability and open software interfaces. OpenArm takes the opposite approach — the mechanical design prioritizes back-drivability and transparency of joint torque so that operators can physically guide the arm during leader-follower teleoperation without fighting against servo resistance.

The "open-source" designation refers to both the hardware and software. The mechanical design files (URDF, STEP, STL) are published, the ROS2 packages are open-source, and the LeRobot data export pipeline is provided without license restrictions. This means research teams can build their own OpenArm units, modify the design for specific task requirements, and contribute improvements back to the community. SVRC sells pre-assembled, tested units for teams that want to start immediately rather than manufacture their own.

OpenArm is the arm used in SVRC's DK1 bimanual kit — the leader-follower teleoperation platform referenced in our ALOHA guide as a lower-cost alternative to the ViperX/WidowX-based ALOHA. Pairs of OpenArm 101 units can be configured as leader-follower pairs for bimanual data collection. The arm is also used as a standalone manipulator for single-arm imitation learning research, where it competes directly with SO-ARM100, Koch v1.1, and ViperX 300.

Full Specifications

The OpenArm 101 specifications reflect its design priorities: enough payload and reach for tabletop manipulation tasks, low joint friction for back-drivable teleoperation, and hardware interfaces that connect directly to ROS2 without requiring proprietary middleware.

The 1kg payload spec is the key differentiator relative to lower-cost arms in the same open-source category. Most SO-ARM100 and Koch v1.1 builds have effective payloads of 200-500g at their useful reach, which constrains the range of manipulation tasks that can be collected. OpenArm's 1kg payload at the wrist enables grasping and moving common objects — bottles, mugs, tools, packaged goods — without hitting the payload limit during normal data collection.

OpenArm vs SO-ARM100 vs ViperX 300

Researchers evaluating the OpenArm 101 typically compare it against two other platforms: the SO-ARM100 (a very low-cost community-designed arm) and the ViperX 300 (the arm used in the original ALOHA paper from Stanford). Here is a direct feature comparison:

The practical choice between these three platforms comes down to your research goal and timeline. SO-ARM100 is the right choice only for teams that want to experiment with LeRobot's teleoperation pipeline at minimum cost and are willing to spend significant time on assembly and debugging. ViperX 300 is the right choice if you need exact ALOHA compatibility — for example, if you are building on published ACT results or contributing to the ALOHA open-source ecosystem. OpenArm 101 is the right choice for the majority of imitation learning research teams: higher payload than SO-ARM100, native ROS2 and LeRobot support, factory assembly and testing, and same-day availability in the Bay Area.

ROS2 and LeRobot Setup

OpenArm 101 ships with a full ROS2 package that installs on Ubuntu 22.04 with ROS2 Humble or Iron. The package includes a hardware interface node, URDF model, MoveIt2 configuration, and a joint state publisher. The typical setup sequence for a new unit is:

For teleoperation data collection with LeRobot, OpenArm integrates directly with Hugging Face's lerobot library. The OpenArm robot configuration is included in the LeRobot package as a registered robot type. To record episodes:

The data collection produces LeRobotDataset objects stored in HDF5 format with synchronized joint states, actions, and camera frames. The format is compatible with all major VLA training frameworks including ACT, Diffusion Policy, and OpenVLA. See our LeRobot getting started guide for the full training pipeline after data collection.

ROS2 OpenArm Topics and Interfaces

Key ROS2 topics published by the OpenArm hardware interface node:

• /joint_states — 6-DOF joint positions, velocities, and efforts at 50Hz
• /openarm/gripper_state — gripper position and force feedback
• /openarm/end_effector_pose — Cartesian pose of the end effector (TF2)
• /openarm/diagnostics — joint temperature, current, error status

Subscribed topics for control:

• /joint_trajectory_controller/joint_trajectory — standard ROS2 control interface
• /openarm/gripper_cmd — gripper open/close commands

Use Cases: Teleoperation and Imitation Learning

Leader-Follower Teleoperation

The most common use of OpenArm 101 in research is leader-follower teleoperation: one arm is configured as the leader (the operator holds it and moves it by hand), and a second arm mirrors the leader's joint positions as the follower (the arm that interacts with the environment and carries the gripper). This setup allows a human operator to demonstrate manipulation tasks while a synchronized recording captures joint positions, gripper state, and camera observations simultaneously.

OpenArm's low joint friction is critical for this application. Arms with high servo resistance require the operator to exert significant force to move the leader arm, which causes fatigue and unnatural demonstrations. OpenArm is designed for smooth back-drivability so operators can provide natural demonstrations over extended sessions without ergonomic strain. Our Wuji Hand teleoperation guide covers how to pair OpenArm with the Wuji Hand for dexterous finger-level demonstrations.

Single-Arm Imitation Learning

For single-arm tasks — pick and place, bin sorting, peg insertion, object handover — OpenArm provides a complete platform with 1kg payload and 650mm reach. This covers the large majority of tabletop manipulation tasks in the imitation learning literature. The arm can be mounted on an aluminum extrusion tabletop fixture and run continuously for data collection campaigns. SVRC runs OpenArm units in our Mountain View facility for managed data collection; see our data services page for rates.

Bimanual Configuration (DK1 Kit)

Two OpenArm 101 units configured as a leader-follower bimanual pair form the basis of the SVRC DK1 kit. This configuration is functionally similar to ALOHA but uses OpenArm mechanics instead of ViperX/WidowX arms. The DK1 kit costs $9,000 assembled and can be operational within 2 hours of unboxing, compared to 20-60 hours for a DIY ALOHA build. See the ALOHA guide for a detailed comparison.

Compatible Accessories

OpenArm 101 is designed to work with a specific set of SVRC-supported accessories. Using verified compatible hardware reduces integration time and ensures that data collection pipelines work without custom adapters.

Wuji Hand

The Wuji Hand is a multi-fingered dexterous hand that mounts directly on the OpenArm 101 wrist flange. It provides 5 fingers with individually actuated joints, enabling dexterous grasping tasks that a parallel gripper cannot perform. The Wuji Hand uses the same USB control bus as the arm, so no additional compute is required. See the Wuji Hand teleoperation guide for setup and recommended tasks.

OpenArm Wiring Harness

The OpenArm Wiring Harness is a structured cable management kit that routes power and communication cables cleanly through the arm's link structure. It eliminates the most common assembly failure mode — cables that restrict wrist rotation or get pinched between links during manipulation. The harness is recommended for any deployment where the arm will perform repetitive high-range-of-motion tasks over extended sessions.

Camera Mount Bracket (Intel RealSense)

SVRC offers a wrist-mount camera bracket compatible with Intel RealSense D405 and D435i cameras. The bracket attaches to the OpenArm wrist link and positions the camera to provide a consistent close-up view of the end effector and workspace. Wrist camera data significantly improves policy performance on contact-rich tasks.

Pricing and Where to Buy

OpenArm 101 is available through the SVRC store. In-person pickup is available at SVRC's Mountain View, CA facility (1117 Independence Ave) with same-day availability when stock is on hand — call ahead to confirm. Shipping to US addresses is 3-5 business days. International shipping available; contact us for customs documentation and estimated duties.

Frequently Asked Questions

What is the OpenArm 101?

OpenArm 101 is a 6-DOF open-source robot arm developed by Robotics Center of Silicon Valley (SVRC). It has a 1kg payload, 650mm reach, and is designed for teleoperation and imitation learning data collection. It is ROS2 native and LeRobot compatible, available at SVRC for $5,400.

What is the OpenArm price?

The OpenArm 101 is $5,400 at SVRC. Same-day pickup is available at the Mountain View, CA facility at 1117 Independence Ave. A bimanual leader-follower pair (DK1 kit) is available for $9,000. Accessories including the Wuji Hand and OpenArm Wiring Harness are sold separately.

Is OpenArm compatible with ROS2 and LeRobot?

Yes. OpenArm 101 is ROS2 native with packages for ROS2 Humble and Iron, including MoveIt2 support. It exports data in LeRobot format directly via the lerobot library, making it straightforward to use with Hugging Face's teleoperation and training framework. The arm is listed as a registered robot type in the LeRobot configuration library.

How does OpenArm 101 compare to SO-ARM100 and ViperX 300?

OpenArm 101 sits between SO-ARM100 and ViperX 300 on the cost-capability spectrum. It has higher payload than SO-ARM100 (1kg vs ~250g), ships pre-assembled (no DIY), and has factory support from SVRC. Compared to ViperX 300, it has native LeRobot support and same-day Bay Area pickup, while ViperX 300 has native ALOHA ACT compatibility and slightly longer reach. OpenArm is the better choice for most imitation learning research; ViperX is better only if you need exact ACT/ALOHA paper compatibility.