Skip to content
buildrobots.io

Path 1 of 3 · A · 5 steps

Arm

A 5 or 6 DOF desktop arm. Kinematics, calibration and teleoperation, then your first learned policy.

  1. 01

    Run a robot in simulation

    Load an MJCF model, step physics, read joint state.

    Done whenYou can script a joint trajectory in the MuJoCo notebook and plot the joint angles.

    Read / watch / run

    • MuJoCo Python tutorial notebook

      Google DeepMind · Notebook · free · colab.research.google.com/github/google-deepmind/mujoco/blob

      Load a model, step physics, render, read sensors and contacts, all in a browser notebook.

    • MuJoCo documentation

      Google DeepMind · Docs · free · mujoco.readthedocs.io/en/stable/overview.html

      The physics engine behind most open robot sims: MJCF models, actuators, contacts, sensors.

    BenchNo hardware. Do it before you buy parts; it is cheaper to break things in sim.
  2. 02

    Build the arm

    Print parts, configure servo IDs, assemble, calibrate.

    Done whenLeader and follower arms calibrated; the follower mirrors the leader without jitter.

    Read / watch / run

    • SO-101 setup guide

      Hugging Face LeRobot · Docs · free · huggingface.co/docs/lerobot/so101

      Source parts, configure Feetech servos, assemble and calibrate the SO-101 leader and follower arms.

    • SO-ARM100 / SO-101 hardware repo

      The Robot Studio · Repo · free · github.com/TheRobotStudio/SO-ARM100

      STL files, BOM with vendor links and assembly notes for the open 5-DOF arm plus gripper.

    • Low-cost robot arm (Koch v1.1)

      Alexander Koch · Repo · free · github.com/AlexanderKoch-Koch/low_cost_robot

      A Dynamixel-based leader/follower arm: printed parts, BOM and wiring.

    Practise on · computed

    Buildable robots of this class, easiest first by the difficulty index; robots with a build package listed first.

  3. 03

    Kinematics and planning

    Forward and inverse kinematics, joint limits, trajectory timing.

    Done whenYou can move the tool tip to a Cartesian point and explain the IK solution you used.

    Read / watch / run

    • Modern Robotics: Mechanics, Planning, and Control

      Kevin Lynch & Frank Park, Northwestern · Book · free · hades.mech.northwestern.edu/index.php/Modern_Robotics

      Rigid-body motion, forward and inverse kinematics, dynamics, trajectory generation. Book, videos and code.

    • MoveIt 2 tutorials

      PickNik Robotics · Docs · free · moveit.picknik.ai/main/doc/tutorials/tutorials.html

      Motion planning for arms in ROS 2: planning scenes, collision checking, trajectories.

    • ROS 2 tutorials (Jazzy)

      Open Robotics · Docs · free · docs.ros.org/en/jazzy/Tutorials.html

      Nodes, topics, services, launch files, tf2 and URDF: the plumbing most robot software sits on.

    BenchThe robot you built in the build step of this path.
  4. 04

    Record demonstrations

    Teleoperate, record episodes with cameras, publish a dataset.

    Done when50 clean episodes of one task on the Hub in LeRobot format.

    Read / watch / run

    • Imitation learning on real-world robots

      Hugging Face LeRobot · Docs · free · huggingface.co/docs/lerobot/il_robots

      Teleoperate, record episodes, push a dataset to the Hub, train a policy and run it on the arm.

    • Robotics course

      Hugging Face Learn · Course · free · huggingface.co/learn/robotics-course

      Robot learning from the ground up with LeRobot: data, policies, real hardware.

    Open robot datasets on humanoidrobots.training
    BenchThe robot you built in the build step of this path.
  5. 05

    Train and run a policy

    Train ACT or a small VLA on your data, evaluate on the real arm.

    Done whenThe policy completes the task on the arm and you logged its success rate over 10 tries.

    Read / watch / run

    • LeRobot source

      Hugging Face · Repo · free · github.com/huggingface/lerobot

      The code behind the docs: robot drivers, dataset format, ACT, Diffusion Policy, SmolVLA.

    • Robotic Manipulation

      Russ Tedrake, MIT · Book · free · manipulation.mit.edu

      Perception, planning and control for picking things up, with runnable notebooks.

    BenchThe robot you built in the build step of this path.

Resource links checked 2026-09-25.