Humanoid Database

 

Disclaimer: All content presented on this site, including robot profiles, images, and descriptions, is provided for editorial, educational and informational purposes only. Humanoid Press does not claim ownership of any third-party trademarks, designs, or intellectual property featured herein. All rights belong to their respective owners.

 

Reachy 2

type:
type 1
{{variant.name}}:
{{opt.name}}
{{opt.value ? '' : opt.name}}
{{opt.value ? '' : opt.name}}
Description

Summary:

 

Reachy 2 is the world’s first fully open‑source humanoid robot designed specifically for embodied‑AI research, manipulation, and VR teleoperation.  

 

It features dual 7‑DoF bio‑inspired arms, Orbita actuators, stereo 3D vision, and a complete ROS 2 Humble software stack.  

 

Reachy 2 can be controlled via Python, VR teleoperation, or Unity‑based interfaces, making it one of the most accessible research humanoids available.

 

Editorial:

 

Reachy 2 is built for labs, universities, and AI teams that need a reproducible, modifiable, open platform.  

Where most humanoids are closed and proprietary, Pollen Robotics takes the opposite approach:

 

- full open‑source hardware  

- open‑source software  

- open‑source simulation  

- open‑source teleoperation  

 

This makes Reachy 2 ideal for embodied‑AI experiments, robot learning, manipulation research, and human‑robot interaction.  

 

Its VR teleoperation latency (~125 ms) and modular design give researchers a fast path from teleop → dataset → autonomous policy.

 

Specifications:

 

Dimensions & Build

- Height: 166 cm   

- Weight: 50 kg   

- Design Intent: Open‑source bimanual humanoid for AI research  

- Construction:  

  - Dual 7‑DOF arms  

  - Orbita bio‑inspired actuators  

  - Modular, research‑friendly frame  

  - ROS 2‑native architecture  

 

Mobility

- Locomotion: Wheeled mobile base (not bipedal) 

- Capabilities:  

  - Indoor navigation  

  - Stable mobile manipulation  

 

Degrees of Freedom

- Total DOF: 21   

- Arms: 7 DOF each  

- Head/Neck: Multi‑DOF for camera alignment  

- Base: Differential wheeled mobility  

 

Actuation

- Actuator Type: Orbita bio‑inspired actuators  

- Capabilities:  

  - Smooth, human‑like motion  

  - High‑precision joint control  

  - Research‑grade repeatability  

 

Manipulation

- Payload: 3 kg per arm (3 kg total)   

- Capabilities:  

  - High‑precision grasping  

  - Bimanual manipulation  

  - VR‑guided object interaction  

  - Autonomous manipulation via Python/ROS2  

 

Perception & Sensors

- Stereo 3D vision  

- Depth sensors (RGB‑D)   

- Head‑mounted cameras  

- Mobile‑base sensors for navigation  

 

Compute & AI

- Software Stack:  

  - Full ROS 2 Humble environment  

  - Python SDK  

  - Unity VR apps  

  - One‑line Docker simulation environment   

 

- Capabilities:  

  - VR teleoperation (~125 ms latency)   

  - Autonomous behaviors  

  - Data collection for embodied‑AI  

  - Sim‑to‑Real workflows  

 

Battery & Power

- Runtime: 8 hours (published spec)   

- Charging: Standard mobile‑base charging  

 

Connectivity

- ROS 2 communication  

- Python API  

- VR teleoperation interface  

- Unity integration  

 

Applications

- Embodied‑AI research  

- Robotic manipulation  

- VR teleoperation studies  

- Human‑robot interaction  

- Academic robotics labs  

- Industrial R&D prototyping  

 

Pricing

- $78,000

 

Pollen Robotics has NOT published:  

- Torque curves  

- Joint‑speed specs  

- Environmental ratings  

- Battery capacity details  

- API performance benchmarks  

 

Image: @PROROBOTS

 

Company Source:

 

https://www.pollen-robotics.com/

 

https://youtube.com/shorts/4JQYY7bkl9Q?si=VbgCxIGVPzYyXOTR

 

Key News & Coverage:

 

https://youtube.com/shorts/Cwe6tGD9VXY?si=o-XKbnvR5FtsDKYZ