KORTHOS / ECOSYSTEM / PRODUCT

Reachy 2

About

About Reachy 2

Reachy 2 combines two seven-axis arms, expressive head movement and a mobile base with lidar navigation. Each arm carries approximately three kilograms, while stereo and RGB-D sensing support object perception, teleoperation and bimanual manipulation across a height-adjustable workspace.

Python and ROS 2 interfaces expose navigation, arm, head and gripper control, and the open-source stack supports integration with external AI models. Virtual-reality teleoperation provides human demonstrations and remote control for data collection. Reachy 2 is commercially available to research laboratories and already operates within major AI-development programmes.

Capabilities

Source-backed capabilities

  • Control modes: Teleoperation
  • Object manipulation

    Claimed - 2025-10-30 - Gemini 336 supplies stereo depth data for Reachy 2 to localise and manipulate objects such as fresh produce.

    Source
  • Object manipulation

    Claimed - Remotely grasp and move real-world objects using VR hand controls and the grippers.

    Source

Supply

Supply / Components

  • Reachy 2

    Oak Ffc 4p Vision Controller - Perception - Luxonis - Pollen Robotics Reachy 2 uses a Luxonis OAK-FFC 4P controller for its stereo head cameras - source: luxonis.com

    Source
  • Reachy 2

    Rgbd Camera - Perception - Orbbec - Pollen Robotics Reachy 2 uses an Orbbec Gemini 336 RGB-D camera in its torso - source: orbbec.com

    Source
  • Reachy 2

    Dynamixel Xm430 Actuator For Reachy 2 Gripper - Actuation - ROBOTIS - ROBOTIS supplies dynamixel xm430 actuator for reachy 2 gripper for Pollen Robotics Reachy 2 - source: docs.pollen-robotics.com

    Source
  • Reachy 2

    Rplidar - Perception - SLAMTEC - Reachy 2 mobile-base specifications identify a SLAMTEC RPLIDAR sensor - source: pollen-robotics.com

    Source
  • Reachy 2

    Bedrock V3000 Industrial Pc - Compute & Control - SolidRun - Pollen Robotics Reachy 2 runs its robot stack on a SolidRun Bedrock V3000 industrial PC - source: docs.pollen-robotics.com

    Source
  • Reachy 2

    Tof Depth Module - Perception - Luxonis - Pollen Robotics Reachy 2 uses a Luxonis OAK-FFC ToF 33D module between its head cameras - source: docs.pollen-robotics.com

    Source
  • Reachy 2

    Arducam Imx296 Rgb Camera Modules - Perception - Arducam - Pollen Robotics Reachy 2 uses Arducam IMX296 RGB cameras with fisheye lenses in its head - source: docs.pollen-robotics.com

    Source

Research

  • Reachy 2 integration in LeRobot

    2025-10-24 - Research Publication - LeRobot documentation describes Reachy 2 setup, teleoperation, data recording and policy workflows. The October 2025 v0.4.0 release introduced plugin-based support for both real robot control and simulation.

    Source
  • Reachy 2 Python SDK and control examples

    Research Publication - Pollen Robotics publishes Reachy 2’s Python SDK with arm, gripper, head, camera, audio and mobile-base examples. The source and notebooks provide reusable interfaces for developing and testing applications on the exact Reachy 2 platform.

    Source

Resources

Media library

Official exact-product documentation, Apache-licensed software and first-party media for Reachy 2. Broken legacy datasheet links are excluded.

Timeline

  • Orbbec Gemini 336 powers Reachy 2 perception

    2025-10-30 - Capability - Gemini 336 supplies stereo depth data for Reachy 2 to localise and manipulate objects such as fresh produce. Orbbec’s MX6800 depth engine combines active and passive imaging for operation across varied lighting conditions. Reachy 2 was already operating in at least eight research laboratories, including Cornell, Carnegie Mellon and CNRS.

    Source
  • Pollen Robotics plans Reachy 2’s CES 2025 debut

    2024-12-09 - Market Signal - The company also planned an appearance at CES Unveiled Las Vegas on January 5, ahead of the main exhibition. Pollen described a global installed base of more than 100 Reachy robots across 20 countries, spanning research laboratories and other organisations.

    Source

Context

Market context

Open and modular mobile manipulator from Pollen Robotics for teleoperation, embodied AI research, and human-facing manipulation without requiring teams to build the full hardware stack.

  • Target environment: Factories, warehouses, labs, service sites and controlled indoor environments where wheeled manipulation can operate safely.
  • Workflow context: Autonomous navigation, object handling, dual-arm or arm-based manipulation, teleoperation and facility task execution.
  • Customer context: Industrial operators, logistics teams, research labs and service organizations evaluating mobile manipulation.
  • Deployment model: Commercial or available product positioned for customer use, integration or developer adoption.
  • Commercial maturity: Publicly presented as commercial or available, with maturity depending on deployment scale and support coverage.
  • Adoption constraints: Adoption depends on site fit, manipulation reliability, safety, navigation, integration effort and task economics.
  • Market position: Reachy 2 sits in the mobile-manipulation lane, using mobility plus arms to pursue practical physical tasks.
  • Adjacent products: Kinisi 01, DexForce W1, Stretch RE2, Quanta X2

Comparable Systems