$8,950Westwood store price (list was $15,290)
AvailableRL walkingUCLA RoMeLa designProprioceptive actuatorsUS-made
Possibly ≤ 1 weekShipping: ~5 business days. Sold directly from Westwood's online store, listed in stock with Add to cart open on an October 1, 2026 check. Its policy says in-stock orders typically ship within 5 business days, and US orders on the USA Air Special Line are delivered duty paid. Westwood store ↗
Overview
Born in UCLA's RoMeLa lab (the group behind DARwIn-OP) and commercialized by Westwood Robotics, BRUCE is a carbon-fiber kid-size humanoid with liquid-cooled knee actuators built for dynamic locomotion research. It is one of the few truly dynamic small humanoids sold by a US company with open control code. The arms are simple, so it is a locomotion platform rather than a manipulator.
Specifications
| Category | Bipeds · Kid-size humanoids (≤ 1.3 m) |
|---|---|
| Released | 2022 |
| Country | USA |
| Availability | Available |
| Programmability | Open source (hardware + software) |
| Built-in autonomy | None / scripted · Moves only when controlled, or plays preset motions |
| LLM support | Possible via SDK · Open control code with SSH access, no LLM features. |
| MuJoCo simulation | Official model · Westwood publishes a BRUCE MJCF and scene. BRUCE_simulation_models ↗ |
| NVIDIA Isaac simulation | Maker URDF only · Westwood publishes a BRUCE URDF, no Isaac package. BRUCE_simulation_models ↗ |
| Gazebo simulation | Sim-to-real kit · BRUCE-OP's Gazebo simulation runs the same control code as the real robot (not ROS). BRUCE-OP ↗ |
| Best for | Professionals & labs |
| Vision sensors | No vision sensors |
| Actuators | Brushless / direct-drive motors |
| Audio | No audio |
| Compute type | Rockchip |
| Balance type | RL walking |
| Height | 70 cm |
| Weight | 4.8 kg |
| Compute | Khadas Edge2 (Rockchip RK3588S, ~6 TOPS), 8 GB RAM, 32 GB storage |
| Speed | Not published |
| DOF | 16 (5 per leg, 3 per arm) |
| Sensors | 6-DoF IMU (up to 2 kHz), foot contact sensors. 14.8 V 3 Ah battery, about 20 min runtime |
| Balance | Proprioceptive, backdrivable actuators with model-predictive and learned controllers. Walks, runs and jumps dynamically. |
| SDK | Open platform: BRUCE-OP GitHub code, SSH access, standard deep-learning frameworks |
Ratings & reviews
Ratings and reviews load on the live site.
Booster K1
Mini Pi+
Bumi
R1 (Air / R1 / EDU)
N2
SA01