Python API

The Dropbear Python API provides a high-level interface for controlling the robot, reading sensors, and deploying learned behaviors — without writing ROS 2 boilerplate. It's designed for researchers, builders, and anyone who wants to get a Dropbear doing things fast.

Installation

pip install dropbear-sdk

# Or install from source
git clone https://github.com/Hyperspawn/Dropbear.git
cd Dropbear
pip install -e ".[dev]"

Quick Start

Connect to a running Dropbear (real or simulated) and send your first command:

from dropbear import Robot

# Connect to the robot (auto-discovers via ROS 2)
robot = Robot()

# Read current joint positions
joints = robot.get_joint_positions()
print(f"Current joints: {joints}")

# Command the robot to stand
robot.stand()

# Move the right arm to a target pose
robot.move_arm("right", position=[0.4, -0.2, 0.8], orientation="forward")

# Walk forward 1 meter
robot.walk(distance=1.0, speed=0.3)

Core API: Robot Class

The Robot class is the primary interface. It wraps ROS 2 communication and provides synchronous methods.

Sensor Access

# Camera stream
frame = robot.get_camera("head")  # Returns numpy array (H, W, 3)

# Force/torque sensors (feet)
left_ft = robot.get_foot_force("left")
right_ft = robot.get_foot_force("right")

# Battery status
battery = robot.get_battery()
print(f"Battery: {battery.percentage}% ({battery.voltage}V)")

Skills Integration

Deploy learned behaviors directly from the Hyperspawn Skills Marketplace:

# List available skills
skills = robot.list_skills()

# Deploy a walking policy trained in Isaac Lab
handle = robot.deploy_skill("bipedal-walk-v2")
handle.start()

# Monitor execution
while handle.running:
    print(f"Status: {handle.status}")
    time.sleep(1)

handle.stop()