Unitree Go2
Unitree consumer/educational quadruped robot with 4D LiDAR. 15 kg, 3.7-5 m/s top speed, GPT integration. Air/Pro/EDU variants from USD 1,600.
Applications
- Research
- Education
- Inspection
- Safety
- Entertainment and events
Unitree SDK2 (unitree_sdk2) is the second generation official developer kit from Unitree Robotics — the manufacturer of the most popular and widely available quadruped and humanoid robots on the market. SDK2 replaced the older Unitree Legged SDK and is a single, unified codebase for all currently produced platforms: Go2 (consumer/pro quadruped), B2 (industrial quadruped), H1 (research humanoid), G1 (advanced humanoid), and robotic arms Z1/B2W.
The communication foundation of Unitree SDK2 is CycloneDDS (DDS middleware compliant with the DDS/RTPS specification), providing deterministic, publish/subscribe communication over WiFi or Ethernet. The SDK exposes a complete API for: reading sensor states (IMU, proprioception 12-20 DOF, RGB/depth cameras), sending control commands (torque, position, velocity), calling predefined behaviours (standing, walking, trot, gallop, running, step climbing), and managing the payload (AI-enable 3D LiDAR, Z1 manipulation arm).
SDK2 is written in C++17 with complete Python 3.8+ bindings. The official unitree_ros2 package provides ROS 2 layers (Humble, Jazzy): robot state topics, velocity control via cmd_vel, Nav2 integration for autonomous navigation. The Unitree Go2 with AI Kit (Jetson Orin NX + Hesai 4D LiDAR + IMU sensors) is a ready-to-use system for autonomous exploration, mapping and environment inspection.
In the robot learning ecosystem, Unitree Go2/G1 are the most widely used research platforms in the world — low price (Go2 Pro ~$3,500, G1 ~$16,000), good documentation and SDK2 with ROS 2 make them the de facto standard in academic robotics labs. Tens of thousands of Go2 units are used in academic settings for research on locomotion, reinforcement learning and sim-to-real transfer.
Unitree SDK2 is open-source under BSD 3-Clause, with a regular release cycle (~every 2-3 months) and active support from Unitree via GitHub Issues and a dedicated Discord.
An API Library is a software package that exposes programmatic interfaces for communicating with a device, service, or system. In robotics it typically forms a lightweight integration layer built on top of the manufacturer's official API or an open-source project, abstracting low-level protocol details and providing language-native bindings (Python, C++, Java, etc.).
A Control Stack is the set of software components responsible for control logic, motion planning, command execution, and coordination of a robot's actuators. It receives high-level goals or waypoints (from planning or AI layers) and translates them into low-level actuator commands, closed-loop feedback control, and safety checks. Common implementations include MoveIt (arm manipulation), Nav2 (mobile navigation), and custom whole-body control stacks for humanoids.
An SDK (Software Development Kit) is a curated set of libraries, interfaces, tools, sample code, and documentation intended for building applications and integrating with a specific hardware device, platform, or service. In robotics, an SDK typically exposes device control, telemetry, sensor access, configuration, and execution functions, significantly reducing the time-to-first-integration for developers targeting a specific robot or platform.
Robot Control denotes the role of software responsible for motion control, command execution, coordination of actuating elements and the direct operational logic of the robot.
The robot learning role describes software for training a robot's control policies and manipulation/locomotion skills using machine learning methods. It covers: reinforcement learning in simulation with sim-to-real transfer, imitation learning and learning from demonstration, training Vision-Language-Action (VLA) models, and fine-tuning robotics foundation models. It typically uses massively parallelized simulation environments (Isaac Lab, MuJoCo) to generate training data, then deploys the trained policies on a physical robot.
Device Integration denotes the role of software responsible for communication, configuration, initialization and handling of specific devices, sensors, controllers or hardware components within a robotic system.
Global academia — over 200 research groups worldwide use Unitree Go2 + SDK2 for research on locomotion, reinforcement learning, sim-to-real and SLAM. ETH Zurich, MIT CSAIL, Stanford, CMU, Berkeley, Oxford, TU Berlin, KAIST and many more. NVIDIA GR00T / Isaac Lab — Unitree G1 is an officially supported platform for NVIDIA Isaac Lab and the GR00T N1 fine-tuning pipeline. Unitree and NVIDIA announced a partnership at GTC 2025. Hugging Face LeRobot — Unitree Go2 is among the platforms supported by LeRobot for demonstration data collection and robot learning. Infrastructure inspection — Go2 with AI Kit (LiDAR + camera + autonomy) deployed by several inspection companies for industrial facility, tunnel and power plant inspections. Military and defence (research) — US Department of Defense (DARPA, Army Research Lab) purchased dozens of Unitree Go1/Go2 units for research on locomotion in difficult terrain. Parkour and acrobatics demonstrations — viral videos of Go1/Go2 performing backflips, parkour and dancing; hundreds of millions of views on social media, building brand recognition.
GitHub: unitreerobotics/unitree_sdk2 — ~1 500 stars, ~350 forks, ~20 contributors. Active issues, regular releases. Ecosystem — dozens of related repositories: unitree_ros2 (~700★), unitree_mujoco (~500★), unitree_isaac_sim, go2_ros2_bridge and others. Unitree Developers Discord — official server with >5,000 members, real-time support. Academic citations — Unitree Go2 appears in hundreds of conference papers (ICRA, IROS, CoRL, NeurIPS, ICLR) as an experimental platform. Global distributor network — 30+ official distributors in Europe, America and Asia, providing local support for academia and industry. Dominance in the affordable segment — Unitree holds over 60% of the quadruped market in the <$20,000 price category, translating to the largest SDK2 user base among quadruped platforms.
Unitree consumer/educational quadruped robot with 4D LiDAR. 15 kg, 3.7-5 m/s top speed, GPT integration. Air/Pro/EDU variants from USD 1,600.
Unitree industrial quadruped robot with IP67 rating. 60 kg, 40 kg continuous / 120 kg static payload, 6 m/s top speed, 12 DOF.
Bipedal humanoid robot by Unitree Robotics, designed as a compact research, development, and developer platform.
Full-size general-purpose humanoid (~180 cm, ~47 kg). Bipedal, 5 DOF per leg, 3.3 m/s walking speed, 360° perception via 3D LiDAR + depth camera, Unitree M107 PMSM joint motors (~360 N·m peak knee torque).
Ubuntu 24.04 LTS 'Noble Numbat' — supported until April 2029. The host for ROS 2 Jazzy.
Communication: WiFi 5 GHz or 100M/1G Ethernet. CycloneDDS requires network interface configuration. DDS security recommended for field deployment. SDK2 supports multiple simultaneous clients.
License family: Permissive
Unitree G1 (humanoid) support, full whole-body kinematics, new G1 arms API.
Go2 Air/Pro AI Kit support, Hesai 4D LiDAR integration, DDS stability improvements.
Refactored to CycloneDDS, unified API for Go2/B2/H1/G1, Python 3.8+ bindings.
unitree_ros2 package for Humble and Jazzy, Nav2 integration, Gazebo Harmonic support.
First public SDK2 release for Go2. Replaced Unitree Legged SDK.
SDK2 beta for partners — basic Go2 gait API over CycloneDDS.
