The Printed Quadruped
The iconic Physical AI form: a twelve-servo legged dog you print at home. A central-pattern-generator turns a stick into a stable trot, the same gait you'll port to the servos.
Drive it in real physics.
The four-part build starts with this: the design as a real MuJoCo twin you can pilot right now on Play. Take the controls, then build the body that runs it.
Walk the twin
The printed quadruped in real MuJoCo physics. Drive it with a controller or the keyboard; a central-pattern-generator holds the trot while you set the direction.
Gamepad or keyboard: steer, turn, and watch the world it maps through its rangefinder.
▶ Drive the twin on Play ↗Four parts, one loop.
Body, brain, driver, policy. The design forces no single board, only that whatever you pick can take the commands and run the policy.
A twelve-servo legged body
Four three-joint legs on a central spine that carries the board and battery, printed in place where the design allows. Hobby servos, no machined parts.
Whatever fits your budget
Twelve joint channels and a command link, a servo bus plus a small board. Pick from the options below; the design locks to none of them.
Browser → servo bus
Twelve joint angles per tick stream to the servos over the driver registry, USB-serial to a servo controller, or straight to smart servos. The gait runs on your laptop, the legs on the body.
A gait you can grow
A central-pattern-generator maps your heading command to twelve leg angles. Start from Forge's open-loop trot; graduate to a gait that learns to walk on-device.
Ways to power the same body.
Cheapest to heaviest. Every one drives the identical printed body and runs the same policy.
Servo driver + Espressif ESP32-S3
~$20Drives twelve servos from a bus and takes gait commands from your laptop over serial. The cheapest way to make the legs move.
Raspberry Pi 5
$80+Runs the gait and light perception on the dog itself, untethered.
NVIDIA Jetson Orin Nano
$249Learned locomotion and vision on-board: the research path for a walking policy.
Full specs and roles for every board: Digital Brains →
Open robots like this one.
The Maker Atlas maps real, open Physical-AI robots to each of our builds. These are kin to this one, open a profile to trace its boards, sensors, and source.
Petoi Bittle (OpenCat)
The top open quadruped framework, palm-size, Arduino/ESP32 (BiBoard), MPU6050 IMU + an on-module AI vision camera. OpenCat MIT.
In the wildOpenDog V3
James Bruton's large Spot-style open printed quadruped, BLDC via ODrive + 3D-printed cycloidal gearboxes. Open, ~$2k BOM.
In the wildMangDang Mini Pupper 2
A CM4-based education quadruped with an OAK-D depth camera + ROS2 SLAM/Nav2; Stanford-Pupper lineage. ~$599, open.
In the wildSpotMicro
A community SpotMini-style printed quadruped, Raspberry Pi + servos, MPU6050 IMU, ROS inverse-kinematics gait. MIT, widely forked.
In the wildStanford Pupper
The Stanford Robotics Club's open legged-locomotion quadruped, Raspberry Pi + Teensy, the base under Mini Pupper. MIT.
See how they all connect: the Maker Atlas →
Live now: the twin. In progress: the physical unit.
The sim twin on this page is real and live today. The printable body and the driver are specced against the sim-to-real registry; the first physical unit is being assembled, and the STL pack and step-by-step land here the moment it's verified on real hardware, not before.
Design its body, program its brain.
Take this build into the Studio: model and export its printed parts on the CAD kernel, and write the MicroPython that runs on its board. Both open as full-screen workspaces.