Microduck is a 25cm, $399 open-source bipedal robot built by Hugging Face and Pollen Robotics. It’s designed for sim-to-real reinforcement learning. It ships with 7 pre-trained behaviors and an Apache 2.0 software stack you can clone, modify, and retrain.

Payment Required

Launch tags:Open Source•Robots•Artificial Intelligence

Launch Team

SkydiveAI coworkers that actually get work done

Promoted

Love that you made a real bipedal robot hackable under Apache 2.0, turning sim-to-real RL from a lab luxury into something anyone can retrain on their desk.

Report

1d ago

1 2

Previous Hugging Face Launches

MorphAISeamless Facial Identity Transformation

Launched on February 18th, 2026

private ai planner for daily tasksprivate ai planner for daily tasks

Launched on February 16th, 2026

MindsEye Dataset Admin Console"The first public dataset of real organizational AI memory

Launched on January 7th, 2026

Kindle Highlight OrganizerTurn Kindle highlights into usable knowledge

Launched on December 27th, 2025

View all Hugging Face launches

5.0

Based on 92 reviews

Review Hugging Face?

Reviewers largely see Hugging Face as the default hub for open-source AI: a practical place to find, compare, test, host, and share models and datasets, with strong tooling like Transformers and Spaces that speed experimentation. Users and founders alike praise its breadth, documentation, and community pull. Makers of Inrō, IFTTT, and Typeahead specifically cite its model ecosystem, inference endpoints, and compatibility. The main complaint is usability: newcomers can struggle with onboarding, model selection, and uneven repository quality, licensing, and documentation.

+89

Summarized with AI

Pros

open source models (14)

AI community (8)

model hub (6)

ease of use (5)

AI model repository (4)

community support (4)

open datasets (4)

collaboration on AI models (3)

inference endpoints (3)

dataset curation (2)

Cons

computationally heavy models (2)

Pros

Cons

Reviews

All Reviews

Most Informative

•16 reviews

The models perform well, and the Transformers library is robust, although some models are computationally heavy, which can slow down smaller projects. Still, it’s a solid tool.

What's great

Transformers library (2)

What needs improvement

computationally heavy models (2)

Report

45 views2yr ago

•9 reviews

✨ As a new profile working in product and operations, I genuinely admire what Hugging Face is building. You've made AI feel more accessible, collaborative, and... actually understandable! I might not write the models, but I can explore, learn, and even imagine use cases thanks to the clarity and openness of your platform. This is exactly the kind of shift that empowers cross-functional teams to innovate together. 🚀 #AIForEveryone #ProductThinking #FutureOfWork

What's great

ease of use (5)AI community (8)collaboration on AI models (3)

Report

83 views1yr ago

used Hugging Face to build Aloud

•4 reviews

transformers.js runs Whisper ONNX inside Electron on WebGPU, which is the only reason Aloud can transcribe without sending anything anywhere. A cloud speech API would have been a day of work instead of a week — and it would have cost me the whole promise. My privacy page says audio and video never leave your Mac, and that sentence is true because of this.

Alternatives Considered

Report

9 views10d ago