Open-Source Toy Robots

Hugging Face Launches Its Duck Robot

Hugging Face launched the Microduck, a 25-centimeter open-source duck robot aimed at developers and hobbyists, featuring a camera, lidar and dual IMUs and designed to be trained with reinforcement learning. The company introduced Microduck through Pollen Robotics, which Hugging Face acquired in 2025, positioning the robot as a hands-on platform for physical AI experimentation.

The duck can waddle, pick up objects with its beak, right itself after falls, crouch and even roller skate, with behaviors that can be trained in simulation and then deployed to the device. An SDK, simulation environment and full reinforcement learning training stack are available on GitHub, allowing developers to fine-tune, retrain and redeploy behaviors.

For developers, Microduck lowers the barrier to experimenting with embodied AI and robotics by offering a $399 open-source platform for prototyping real-world behaviors. Its open-source design emphasizes developer control and auditability while also highlighting privacy considerations when third-party applications access onboard sensors.

Image Credit: Hugging Face

Open-source Robotics
Affordable hardware, shared software stacks and transparent designs create new space for rapid experimentation in embodied AI and developer-led automation.
Embodied AI Prototyping
Simulation-trained behaviors that transfer to physical devices make robotics development more accessible for hobbyists, startups and enterprise innovation teams.
Reinforcement Learning Toys
Playful robot formats turn advanced AI training methods into approachable tools for education, prototyping and consumer technology exploration.

Who This Affects Most

Consumer Robotics
Low-cost programmable robots signal a shift toward customizable household and hobbyist devices that evolve through community-built behaviors.
AI Development Tools
Integrated SDKs, simulation environments and training pipelines expand the market for platforms that bridge software engineering and real-world machine intelligence.
STEM Education
Hands-on robotic kits with sensors and AI training capabilities create richer learning environments for coding, machine learning and physical computing.
SCORE
3.7 out of 10
GENDER
50% Men50% Women
MARKETTop markets: North America, Europe, Asia
GENERATION
  • Gen Alpha
  • Gen Z (primary audience)
  • Millennial (primary audience)
  • Gen X (primary audience)
POPULARITY
Popularity 11%
Activity 0%
Freshness 100%