Contact Sales
Back to Learn

OpenArm 1: From Unboxing to Your First Trained Policy

The complete OpenArm learning path. In ~12 hours go from unboxing your robot to training and deploying your first imitation learning policy. Beginner-friendly, hardware-first.

A structured, sequential path that takes you from zero to running a real imitation learning policy on your OpenArm. No robotics experience required — only Linux basics and Python.

Total Time ~12 hours

Difficulty Beginner-friendly

Hardware OpenArm 1 + USB-C cable

Prerequisites Linux basics, Python basics

Simulation Option Yes — sim setup guide

You Will Build A pick-and-place demo with a trained policy

Before You Start — Check These Prerequisites

  • Comfortable with a Linux terminal (cd, ls, pip install)
  • Python 3.10 or higher installed
  • Access to an OpenArm 1 or a simulation environment
  • A machine running Ubuntu 22.04 or 24.04 (VM is fine)
  • About 12 hours of total time across multiple sessions

Not sure if you qualify? Start with Unit 0 — it exists specifically to answer this question.

Your Path at a Glance

Complete the units in order. Each unit has a clear completion check — don't move on until you pass it.

Before You Begin: What You NeedRequiredHardware checklist, software checklist, time estimates, and where to get help. No technical content — just orientation.~20 min · No code
Get Your OpenArm MovingRequiredUnbox, connect, install SocketCAN and ROS 2, calibrate, and move the arm for the first time via the web UI.~2 hours
Install the SDK and ConnectRequiredInstall the OpenArm Python SDK, run a live connection test, and fix the three most common errors.~1.5 hours
Control the Arm with Your HandsRequiredUnderstand what teleoperation is and why it creates training data. Set up a session and teleoperate for 5 continuous minutes.~2 hours
Record Your First DatasetRequiredLearn why data quality matters more than quantity. Record 50 clean pick-and-place demonstrations in LeRobot format.~2 hours
Train Your First PolicyRequiredTrain an ACT policy on your dataset. Understand the training curves and know when to stop. Target: >70% success rate.~3 hours
Deploy and ImproveRequiredRun your policy on the real arm, evaluate it systematically, and start the data flywheel. Target: 7/10 autonomous success rate.~1.5 hours

Time Breakdown

Unit Topic Time
Unit 0 Orientation ~20 min
Unit 1 Hardware Setup ~2 h
Unit 2 SDK & Connection ~1.5 h
Unit 3 First Teleoperation ~2 h
Unit 4 Data Collection ~2 h
Unit 5 Policy Training ~3 h
Unit 6 Deploy & Improve ~1.5 h
Total ~12 h 20 min

Start Unit 0 — Orientation →

Takes about 20 minutes. No technical content — just makes sure you have everything before the path begins.