VLAI L1
A high-performance mobile dual-arm robot with 16 DOF, ±0.02mm repeat positioning accuracy, and native ROS2 + MoveIt2 -- built for manipulation research, imitation learning, and production teleoperation. US sales and support by RoboticsCenter.
Full Hardware Specs
VLA Training Pipeline
The Developer Pro and Developer Max tiers include onboard compute (70-128 TOPS) and a one-click pipeline for training vision-language-action (VLA) models from teleoperation demonstrations.
Supported Architectures
ACT (Action Chunking Transformer), Diffusion Policy, OpenVLA, and RT-2 style models. The pipeline handles data format conversion, training, and checkpoint export.
Data Collection → Training
Record episodes via browser teleop or VR controller (Pro/Max). Episodes auto-upload to the SVRC data platform. One-click training launches on the onboard GPU or cloud.
Benchmark Tasks
VLAI provides reference benchmarks for pick-and-place (95%+ success), drawer open/close (90%+), and kitchen manipulation (85%+) trained from 50-200 demonstrations.
SVRC Data Platform Integration
The VLAI L1 integrates natively with the SVRC Fearless Platform for data collection, model training, and deployment.
Session-Based Recording
Create a session, register the L1 as a device, and start recording. Joint state, camera, and gripper data stream at 30 Hz into JSONL episodes with synchronized timestamps.
Episode Browser
Replay any demonstration in the web-based episode browser. Scrub through joint trajectories, view camera feeds, and annotate task segments for training data curation.
One-Click Deploy
After training, deploy a policy checkpoint to the L1's onboard compute. The platform handles model packaging, inference optimization, and real-time control loop setup.
Use Case Examples
Bimanual Pick-and-Place
Both arms coordinate to pick large or awkward objects. The 6 kg per-arm payload handles boxes, trays, and lab equipment. Trained policies achieve 95%+ success from 50 demonstrations.
Kitchen / Household Tasks
Open drawers, pour liquids, sort utensils. The 106-162 cm lift range reaches countertops and upper cabinets. VR teleoperation (Pro/Max) enables natural demonstration collection.
Lab Automation
Pipetting, plate handling, and sample transport. The ±0.02mm accuracy ensures repeatable positioning for precision tasks. Mobile base moves between workstations autonomously.
Getting Started with the L1
Three steps to your first teleoperation session.
Choose Your Tier
All tiers share the same mechanical frame, grippers, and dual-arm kinematics. The differentiators are onboard compute, bus protocol, bundled software, and warranty.
Typical US lead time: 6-8 weeks. Leasing available on all tiers. Contact RoboticsCenter for final quotes and ordering.
Your Setup Journey
From delivery to first teleoperation session -- estimated 2-3 hours.
Assembly & Inspection
Unpack, inspect dual arms and gripper mounts, verify mobile base wheels
SDK Installation & Network Setup
pip install roboticscenter, connect L1 to local network, rc connect --device l1
Dual-Arm Calibration
ROS2 joint calibration, MoveIt2 home position, gripper calibration
Mobile Base Navigation
WASD test drive, lift range test (106cm-162cm), speed calibration
Manipulation Test
First pick-and-place task, MoveIt2 motion planning, gripper force test
Data Collection & Teleop
Browser teleop panel, session recording, one-click data pipeline
Services for VLAI L1
Data Collection Service
SVRC operators collect dual-arm manipulation demonstrations at our Mountain View or Allston labs. You receive cleaned, platform-ready JSONL episodes.
Hardware Leasing
Lease the L1 on a monthly basis. All tiers available. Includes SDK access, firmware updates, and priority support.
Repair & Warranty
US-based warranty service through RoboticsCenter. Gripper replacement, joint calibration, and mobile base maintenance.
Community
Questions about the VLAI L1 or want to share your manipulation research?
VLAI L1 Community → Ask a Question →