MicroDuck

Microduck reviews: what reviewers and owners actually say

An honest roundup of Microduck reviews: launch coverage, GitHub activity, the community to ask, and a five-point checklist for judging any Microduck review you read - including how to review one yourself.
Oct 13, 2026

A Microduck robot walking across a desk in a warm living room

Full disclosure: this page is written by the team that sells Microduck. So instead of inventing quotes from reviewers we cannot name, this page does something more useful - it shows you where the independent signal actually lives, and gives you a checklist to judge every Microduck review you find. If you searched for "microduck review" or "microduck robot reviews", you will get further with this than with a list of stars.

The independent scoreboard

The strongest review signal for an open-source robot is not a rating out of five, it is whether strangers keep showing up:

  • The GitHub repositories. The robot repository and the RL environments are public - stars, issues and pull requests all tell you how alive the project is, and you can read the code yourself.

  • The Discord. The Pollen Robotics community is where owners compare notes, share policies and post failures. It is the closest thing to unfiltered owner reviews, and nobody there is paid to be nice.

  • The Hugging Face Space. The simulator's activity is another honest heartbeat metric - if the tooling were broken, that Space would be empty.

  • Video. The launch film and community clips show the robot in motion, which is the only "review" a walking robot really needs: does it walk, and does it get back up?

What reviewers tend to agree on

Across coverage and community discussion, the recurring themes are consistent:

  • The motion is the point. Watching a 25 cm robot keep its balance is the moment people understand why this is not a toy - and the demo that consistently converts skeptics.

  • $399 is the headline. Nearly every write-up compares it to Reachy Mini, which starts at the same price (the Lite at $399, Wireless at $499) with a very different mission.

  • Open source is the differentiator. The fact that the walking policies and training environments are public is the thing experienced roboticists single out, because most walking robots are black boxes.

  • The learning curve is real. "It is not a plug-and-play appliance" is a fair summary from the technical side - it ships playable, but its ceiling is where you take it.

How to judge a Microduck review in five checks

  1. Did the reviewer touch the hardware, or only the store page? Hands-on reviews show the robot falling over and recovering. Page-read reviews recycle specs.

  2. Do they show sim-to-real? The interesting claim about Microduck is that sim-trained policies run on real motors. A review that skips that skips the product.

  3. Do they link the repositories? If a review says "open source" without pointing at code, it is marketing, not reviewing.

  4. Do they confuse it with duck toys? Cheap plastic ducks share the name - a review that compares Microduck to a bath toy has reviewed the wrong product. (We wrote the is Microduck a toy? explainer for exactly this confusion.)

  5. Do they get the price right? $399 before taxes and shipping for the robot; packs at $39, $39 and $119. A review quoting a different number is quoting something else.

The questions owners ask before buying

These are the honest, unsponsored concerns that come up again and again - with straight answers:

  • "Is it a toy or a real robot?" A real, open-source robot. It walks, balances, and runs policies you can retrain. Full answer here.

  • "What can it do out of the box?" Seven trained moves: standing up, sitting and standing, walking, turning, kicking, driving and grabbing.

  • "What do I need to know?" Nothing to play with it; Python to program it; persistence to train it. The training guide is the honest picture of the deep end.

  • "Is it worth it against Reachy Mini?" Different robots, same entry price - the comparison lays it out without spin.

  • "What if something breaks?" Email support, and the Dev pack stocks the parts that matter (motors, cables, batteries).

  • "Will more be possible later?" That is the entire design: the community's policies extend the robot, and the repositories are where they land.

Review it yourself

The most valuable review of Microduck is the one you can run without buying anything:

  1. Open the simulator in your browser.

  2. Replay the seven trained moves and try to break the walking.

  3. Read a policy's training environment - the reward function is right there in the repo.

  4. If it holds up in the browser and the code, judge the hardware on the same terms.

When your robot arrives, the review writes itself: film it getting up after a shove, post the policy that failed spectacularly before it worked, and share it in the Discord. That is how reviews work in open source - evidence, not vibes.

Microduck reviews: FAQ

Are there professional reviews of Microduck? Coverage exists from the launch and from robotics channels; it grows with every shipped batch of robots. The repositories and Discord are where the continuously updated, unfiltered version lives.

Can I trust this page? It is written by the seller, so we kept it to verifiable things - repositories, prices, the simulator - and pointed you at independent sources for everything subjective.

Where can I ask real owners questions? The Pollen Robotics Discord. Owners compare policies, prints and falls there daily.

What do owners complain about? The honest recurring theme is the learning curve: it is easy to drive and hard to master. Nobody promised a rubber duck.

Should I wait for reviews before pre-ordering? Try the simulator first - it is free and it is the same robot. That is a better signal than any review, including this one.

Where to go next

Microduck reviews: what reviewers and owners actually say