
Reachy Mini is a compact, open-source humanoid robot platform for research and education, enabling manipulation, vision-based interaction, and reinforcement learning experiments.
Reachy Mini is a compact, open-source humanoid robot designed as an accessible platform for embodied AI research, robotics education, and human-robot interaction experiments. Built around affordable, 3D-printable hardware and off-the-shelf components, it provides a physical testbed for integrating perception, planning, and control models with real-world actuation. The system typically includes an articulated head and upper body with camera-based vision, expressive movements, and programmable behaviors, all controlled via standard software interfaces and modern AI frameworks.
Developers can use Reachy Mini to prototype interactive assistants, study social robotics, or test multimodal models that connect language, vision, and motor control. It supports remote control, scripted sequences, and policy-based control, making it suitable for reinforcement learning, imitation learning, and teleoperation scenarios. Because it is open and modular, researchers can extend the platform with custom sensors, actuators, or AI pipelines, and share reproducible experiments with the broader community.
Please sign in to comment
π¬ No comments yet
Be the first to share your thoughts!
Explore 43+ top alternatives to Reachy Mini

Vention is a cloud-based platform that lets users design, simulate, automate, order, and deploy custom manufacturing equipment directly from a web browser.

Ai2thor is an open-source platform that provides interactive 3D indoor environments and simulation tools for developing, training, and evaluating embodied AI agents.

Project Genie is a web-based experimental research tool that uses world models to let users generate, navigate, and remix interactive environments from text and images.

Sync labs is a platform that lets teams create, manage, and collaborate on AI agents and workflows for data analysis, automation, and application integration.
Depth Anything 3 is a state-of-the-art monocular depth estimation model designed to recover detailed

OpenManus is a framework for training reinforcement learning agents to perform dexterous in-hand manipulation using physics simulation, motion capture data, and real-world robotic hardware.

Realbotix designs modular robotic heads with detachable magnetic faces, enabling rapid, customizable replication of human or humanoid facial structures for lifelike robotic and animatronic applications.

Nvidia Omniverse is a platform for building, simulating, and connecting physically accurate, real-time 3D applications and collaborative virtual worlds using USD-based workflows.