ALOHA / Mobile ALOHA Teleoperation System
A bimanual teleoperation robot based on the Stanford ACT and Mobile ALOHA research platform. Two 7-DOF arms (2x ViperX 300 or compatible) for a total of 14 DOF — the standard rig for household manipulation, imitation learning, and VLA training data collection. SVRC builds and integrates complete ALOHA systems, from $20,000.
Full System Specs
Built for Bimanual Manipulation & Policy Learning
ALOHA was designed from the ground up for collecting high-quality bimanual teleoperation demonstrations. Two arms, one operator, household-scale tasks.
Imitation Learning Data Collection
The original ALOHA paper demonstrated 50-episode ACT policies on tasks like slotting a battery, opening a bag, and cooking shrimp. Leader-follower kinesthetic teaching makes demonstration collection fast — one operator can record 20-30 episodes per hour on practiced tasks.
Household Task Research
Mobile ALOHA expanded the task set to whole-body manipulation: loading a dishwasher, wiping a stovetop, conducting package handoffs. The wheeled base adds navigation to bimanual dexterity, enabling long-horizon household task research.
VLA Training Data
ALOHA is the primary data collection platform for several leading VLA models. Its standardized episode format (HDF5 with joint states + wrist + overhead cameras) is compatible with the ACT training repo, LeRobot, and the SVRC data platform export pipeline.
Works with the SVRC Data Platform for episode recording, QC, and training export.
Stanford ACT & Mobile ALOHA
ALOHA originated from the Stanford Human-Centered AI lab. Understanding the research lineage helps calibrate expectations for what the platform can and cannot do.
Original ALOHA (2023)
Tony Zhao et al. introduced ALOHA with ACT (Action Chunking with Transformers). 50 demonstrations per task were sufficient for high success rates on fine manipulation tasks. The paper established the leader-follower kinesthetic teaching paradigm that the whole field now uses.
Mobile ALOHA (2024)
Zipeng Fu et al. added a wheeled mobile base, extending the platform to whole-body tasks. Co-training on existing ALOHA datasets improved performance. Demonstrated tasks include cooking, cleaning, and social navigation with arm interaction.
ALOHA 2 & Beyond
Google DeepMind collaborated on ALOHA 2, improving the mechanical design for more consistent demos. The ecosystem now includes multiple commercial forks and open-source derivatives. SVRC builds systems compatible with all major variants of the ACT training repo.
Bimanual Teleoperation Comparison
From Demo to Trained Policy
The ALOHA + SVRC platform workflow takes you from hardware setup to policy-ready episodes in a single afternoon.
1. Calibrate & Practice
Calibrate the leader-follower arm correspondence. Practice the target task 5-10 times to build consistent motions. Consistency matters more than speed for demonstration quality.
2. Record Episodes
Use the SVRC platform or ACT data collection scripts to record HDF5 episodes. Each episode captures joint states for both arms, wrist camera images (2x), and overhead camera at 50 Hz. A typical collection run is 50-200 episodes per task.
3. Train & Deploy
Export episodes in ACT, LeRobot, or RLDS format. Train with the ACT repo (typically 3-5 hours on an A100 for 50 episodes). Deploy back to the follower arms via ROS2. SVRC's data platform handles format conversion and QC filtering.
SVRC Integration Process
SVRC handles the full build — you receive a tested, ready-to-use system. Here is what the process looks like.
Scope & Quote
Book a 30-min call to discuss arm choice, camera config, mobile base option, and delivery timeline
Order & Build
SVRC sources arms, fabricates the frame, wires harnesses, and validates joint calibration
Software Setup
Install ROS2, ACT repo, and SVRC data platform. Configure episode recording and camera synchronization
Operator Training
Hands-on session with your team: leader-follower calibration, demo collection workflow, episode QC review
Collect & Train
Start recording episodes. Use SVRC platform for QC, annotation, and export to your training pipeline
Services for ALOHA
Full System Integration — from $20,000
Complete ALOHA system: arms, frame, wiring, calibration, software setup, and operator training. Ships from Mountain View, CA.
Data Collection Service
SVRC operators collect bimanual manipulation demonstrations at our lab. You receive cleaned, ACT-ready HDF5 episodes. Starting at $500 per task (50 episodes).
Demo & Lab Visit
Come to our Mountain View lab to operate the ALOHA system yourself. 30-min demo sessions available — see leader-follower teleop and episode recording live.
Related Guides
Frequently Asked Questions
What is the ALOHA robot?
ALOHA (A Low-cost Open-source Hardware System for Bimanual Teleoperation) is a bimanual robot teleoperation platform developed at Stanford University. It uses two 7-DOF arms totaling 14 DOF, controlled via a matching pair of leader arms through kinesthetic teaching. It was designed to collect high-quality imitation learning demonstrations for training manipulation policies such as ACT (Action Chunking with Transformers).
How much does an ALOHA robot cost?
SVRC integrates complete ALOHA systems starting at $20,000. This includes two ViperX 300 arms (or compatible), the mounting frame, leader arms, wiring, calibration, and software configuration. Contact us or book a 30-min scoping call for a custom quote based on your configuration needs.
Can I buy an ALOHA robot?
Yes. SVRC sells complete ALOHA integration packages starting at $20,000 through our store. SVRC handles the full build — sourcing arms, fabricating the frame, wiring, software setup, and shipping. You can also book a demo session at our Mountain View, CA lab to operate the system before committing.
What is Mobile ALOHA?
Mobile ALOHA is an extension of the original ALOHA system that adds a wheeled mobile base, enabling the bimanual arms to navigate throughout a home or lab. This makes whole-body manipulation possible — tasks that require both locomotion and dexterous arm control, such as loading a dishwasher, wiping a countertop, or opening cabinets. SVRC builds Mobile ALOHA-compatible systems with custom base integrations on request.
Community
Pricing & Availability
Built and shipped from Mountain View, CA. Custom configurations available — mobile base, arm choice, camera setup. Talk to our team for a scoping call.
AgileX PiPER — from $3,999
6-DOF tabletop arm. One of the most affordable research-grade arms available. ROS2 native, Python SDK, CAN bus.