Dropbear
The Dropbear is Hyperspawn's flagship humanoid robot — a 1.8m tall, 32-DOF, fully 3D-printable platform designed for builders, researchers, and developers. It's built to be affordable enough for individuals and capable enough for real-world tasks.
Technical Specifications
| Parameter | Value |
|---|---|
| Height | 1.8 m (5'11") |
| Weight | 45 kg (99 lbs) |
| Degrees of Freedom | 32 DOF |
| Payload per Arm | 5 kg |
| Battery | 48V Li-ion, 2–4 hr runtime |
| Positioning Accuracy | 0.1° |
| Control Loop | 2ms (500 Hz) |
| Body Material | 3D printed PLA / PETG |
| Actuators | Servo + Quasi-Direct Drive (QDD) |
| Compute (Base) | Raspberry Pi 5 |
| Compute (Advanced) | NVIDIA Jetson Orin |
| Software | ROS 2 Humble, on-edge inference |
Sub-Assemblies
The robot is composed of modular sub-assemblies for easier building and maintenance:
Head (6 DOF)
Stewart platform neck with 6 degrees of freedom. Houses stereo cameras (Intel RealSense), IMU, compute module, and speaker/microphone. Pan, tilt, and fine orientation control.
Torso
Central structural core housing the power distribution board, battery, main wiring harness, and carbon fiber spine. Provides mounting points for arms and head.
Arms (6 DOF each)
Each arm has 6 degrees of freedom powered by high-torque QDD actuators. 5 kg payload capacity. Supports parallel gripper or 5-finger dexterous hand end-effectors.
Legs (6 DOF each)
Each leg has 6 DOF with high-torque actuators at hip, knee, and ankle. 4-bar linkage knee design. Feet with integrated force sensors for terrain-adaptive locomotion.
How to Get One
Buy — $9,999
Pre-assembled and tested. Configurable options for hands, compute, intelligence package, and finish. $500 refundable deposit. Ships Q2 2027 (first 500 units).
Configure & Reserve →Build — $1,500–3,000
Build it yourself over 8–12 weekends with our guided platform. AI-assisted sourcing, print planning, step-by-step instructions, and digital twin rehearsal.
Start Building →CAD & Design Files
All CAD files, STLs, and assembly drawings are open-source and available on GitHub. The design is done in Fusion 360 with full parametric history. Community members are encouraged to fork, modify, and contribute improvements.