ALOHA: THE STANFORD ROBOT THAT LEARNS FROM YOU
The bimanual teleoperation rig behind the viral 'housekeeping robot' — teach it a task in about 50 demos and it does it on its own.
by Tony Zhao, Zipeng Fu, Chelsea Finn (Stanford)
RoboticsAI
- difficulty
- ●●●●○
- time
- a serious project
- license
- MIT
- repo
- repo ACTIVE2,284 stars
●●●●○ · a serious project · MIT · 2,284 stars · repo ACTIVE
WHAT YOU’LL NEED
- 3D printer + filament — printable parts — files are in the repo
- Electronic parts — full list with part numbers in the repo BOM
- Dev board / microcontroller — runs the project firmware
- Soldering kit + basic tools — one-time buy, serves every future build
Partner
COMPAREE VERDICT
ALOHA is the project behind that viral clip of a robot cooking and cleaning — and the real thing is a bimanual teleoperation rig from Stanford that learns tasks from human demonstrations. You puppeteer two arms to show it a job, and after roughly fifty demos it performs the task autonomously with surprisingly high success. Set expectations correctly: the main repo is mostly software (ROS teleop and data collection), the physical build rides on commercial Interbotix arms, and the BOM is a Google Doc rather than a full CAD pack. So this is less 'print a robot' and more 'assemble a serious learning platform.' If modern robot learning is what excites you, few projects get you closer, hands-on.
IN THE REPO
NOT IN THE REPO
- —The main repo is mostly SOFTWARE (ROS teleop + data collection)
- —Some STL/CAD is in the aloha2/ folder (mounts, gripper parts, assembly PDF) but it's not a full CAD pack
- —It's built on commercial Interbotix ViperX/WidowX arms — the BOM lives in a Google Doc
- —Training (ACT) is a separate repo
Can I build this?
Build at your own risk. Projects involve tools, electronics and sometimes mains voltage — follow the creator’s safety notes.
Partner · KickstarterHeyGears G1: 10M+ colors and transparent parts in one print, plus UV printing on flat objects. Figures, parts and labels — no painting.
Videos
Mobile ALOHA: Your Housekeeping Robot2:38
The viral teaser that made ALOHA famous.
Gallery
Start here
Navigation into the creator’s own docs — we don’t rewrite the guide, we route you to the source.
- 1.Read the project page (the original ALOHA)
- 2.Watch the demo (why it went viral)
- 3.Read the paper (ALOHA + ACT)
- 4.Get the hardware assembly doc (the Google Doc BOM/assembly)
- 5.Train with ACT (separate training repo)
Resources
Documentation, files and community threads for this build — we link straight to the original sources and never rehost the creator’s files.
Partner · KickstarterCORE POWER BANK S: 0.2" thin, 2.8 oz, 5000 mAh with 22.5W fast charging. Semi-solid battery, carbon fiber, IPX7 waterproof.
KNOWN ISSUES
- The main repo is mostly software — the physical build rides on commercial Interbotix arms, and the BOM is a Google Doc
- The arms have no brakes: they drop when torque is removed — support them during setup
- You need several USB3 cameras; cheap hubs add latency that hurts data collection
- It's ROS1 (noetic) — plan your environment accordingly
Is this the Google DeepMind ALOHA 2?
No — this is the original Stanford ALOHA (and Mobile ALOHA). DeepMind's ALOHA 2 is a separate, later effort.
How little data does it need?
Roughly fifty demonstrations (about ten minutes of teleoperation) can reach high success on a task — that data efficiency is the headline.
Do I need to print the whole robot?
No — it's built on commercial Interbotix arms. You add mounts, cameras and the software; some parts are in the aloha2/ folder.
What does it cost?
Band $$$ — the commercial arms are the big cost. The BOM doc has current figures.
Community builds
No community builds yet — be the first, we feature the best ones.
Discussion1
FROM THE COMPAREE TEAM
ALOHA reaches high task success from roughly fifty human demos — about ten minutes of teleoperation. Is imitation learning at that scale ready for your kitchen, or still a lab trick?
SO-ARM100: The Low-Cost Robot Arm Behind HuggingFace LeRobot
●●●○○ · the cheap, printable way into the same teleoperation idea
OpenArm: A Research-Grade Force-Feedback Robot Arm
●●●●● · a force-feedback arm for the same research niche
XLeRobot: A Household Robot on an IKEA Cart
●●●●○ · a mobile dual-arm home robot you can print
Tony Zhao, Zipeng Fu, Chelsea Finn (Stanford)
A Stanford team open-sourced a low-cost bimanual teleoperation system and the ACT learning method, showing robots can learn real tasks from a handful of human demos.
DISCLAIMER
- Comparee is not the author of the projects featured here. All rights to each project belong to its creator — every page links to the original source, and we never host creators’ files.
- Information is provided without warranty and may become outdated as projects evolve. Prices are indicative bands only — always check the creator’s parts list for current costs.
- Building and operating any project is at your own responsibility. Protective equipment, safe workshop practice and compliance with local regulations are the builder’s responsibility.

CompareeTEAM1mo ago
A detail we didn't expect: the Interbotix arms have no brakes, so the moment torque is cut they just drop — the docs tell you to physically support them during setup. We also noticed the whole physical BOM lives in a Google Doc rather than the repo, which is mostly ROS teleop software. Anyone running this on ROS1 noetic in 2026, or have you ported it?