MagicBot Z1
About
About MagicBot Z1
MagicBot Z1 is a compact bipedal robot built for dynamic walking, balancing and expressive whole-body motion. Its joint system supports agile movement and recovery, while onboard lidar, RGB-D cameras and an inertial unit provide mapping, obstacle detection and spatial awareness. The perception stack supports autonomous route planning and movement through indoor environments.
MagicLab provides a Z1 software-development kit with high-level motion commands and low-level access to individual joints, sensors, robot status and voice functions. Developers use the platform for education, embodied-AI research, exhibitions and commercial interaction applications. Z1 is commercially offered with purchasing, software and secondary-development support.
Capabilities
Source-backed capabilities
- Operating environments: MWC2026 Barcelona demonstration, Intended demonstration spaces, Intended indoor routes
- Control modes: Autonomous
- Bipedal walking
Claimed - Walk with the advertised human-like gait. - environment: Intended demonstration spaces
Source - Navigation and obstacle avoidance
Claimed - Plan routes and avoid obstacles using the advertised navigation algorithms. - environment: Intended indoor routes
Source - Dance performance
Demonstrated - 2026-03-03 - Reproduce the Spring Festival dance and perform Makeba at MWC. - environment: MWC2026 Barcelona demonstration
Source
Research
- fan-ziqi/robot_lab
2024-07-06 - Research Publication - The source registers exact model-specific Isaac Lab environments for tracked quadruped, wheeled-quadruped and humanoid platforms.
Source - MagicBot Z1 SDK
Research Publication - MagicLab publishes C++ and Python development interfaces, command examples and sensor access for MagicBot Z1. The package includes bindings, build integration and network setup for the robot control interface.
Source - MagicBot Z1 robot models
Research Publication - MagicLab releases MagicBot Z1 URDF and MuJoCo descriptions with shared STL meshes, RViz configuration and a viewer launch. The model README enumerates the articulated joint tree and describes 23 actuated joints.
Source - MagicLab Z1 motion retargeting
Research Publication - MagicLab’s GMR branch retargets human BVH and SMPL-X motion to the MagicBot Z1 skeleton through MuJoCo and inverse kinematics. It includes Z1-specific retargeting configurations, visualization, quality checks and motion export scripts.
Source
Timeline
- MagicLab showcases Z1 dance routines at MWC 2026
2026-03-03 - Demonstration - Z1 performed a Spring Festival Gala routine and the dance “Makeba” in synchronized robot presentations. The Barcelona exhibition formed part of MagicLab’s overseas expansion, which the company said covered 27 countries and regions.
Source - MagicLab Z1 wins standing high-jump bronze
2025-08-16 - Award - The bronze-medal result was 0.52 m, behind RobotEra’s gold and Jianghuai Center’s silver. MagicLab says Z1 competed without hardware modification on its commercially offered platform.
Source - MagicLab launches MagicBot Z1 humanoid
2025-07-08 - Product Launch - The base configuration has 24 degrees of freedom, with expansion options up to 50 and optional dexterous hands. The company lists lidar, depth and stereo sensing, alongside developer options for additional computing and arm articulation.
Source
Context
Market context
MagicBot Z1 is a compact bipedal humanoid platform from MagicLab, positioned around high-dynamic movement, onboard sensing and developer-facing embodied AI experimentation.
- Target environment: Research labs, developer programs, indoor demonstrations and early field trials.
- Workflow context: Locomotion testing, balance recovery, perception demos, embodied AI development and general humanoid experimentation.
- Customer context: AI research labs; university robotics programs; embodied AI developers.
- Deployment model: Developer and research-oriented humanoid platform for labs, demonstrations and embodied AI experimentation.
- Commercial maturity: Early commercial — available as a developer platform with limited field deployment data.
- Adoption constraints: Adoption depends on locomotion reliability, useful manipulation skills, developer ecosystem support and safe operation around people.
- Market position: Compact humanoid platform competing in the lower-cost developer and research segment rather than the full industrial humanoid deployment tier.
- Adjacent products: Unitree G1; ENGINEAI PM01; R1.