Robot helper functions and CLI tools for Unitree Go2. This package provides a high-level helper class (Go2RobotHelper) for simplified robot control, a command-line tool for executing robot actions, and a collection of examples and applications.

• Go2RobotHelper: High-level helper class for simplified robot control with automatic connection management, state monitoring, and emergency cleanup
• CLI Tool: go2action command-line tool for executing robot actions
• Examples: Comprehensive examples for audio, video, data channels, and LIDAR
• Apps: Ready-to-use applications for gamepad control, gesture recognition, and visualization

pip install go2tools

Or install from source:

git clone <repository-url>
cd go2tools
pip install -e .

from go2tools.robot_helper import Go2RobotHelper

async def main():
    async with Go2RobotHelper() as robot:
        await robot.ensure_mode("mcf")
        await robot.execute_command("Hello")
        await robot.handstand_sequence()

# List available actions
go2action --list

# Execute an action
go2action StandUp

# Execute with monitoring
go2action Hello --monitor

The Go2RobotHelper class provides:

• Automatic connection management with context manager support
• Built-in state monitoring and display
• Emergency cleanup and safe robot positioning
• Simplified command execution with error handling
• Obstacle detection control and status querying
• Mode switching (MCF/Sport)

See go2tools/robot_helper.py for full documentation.

The go2action command-line tool allows you to execute robot actions directly from the terminal:

go2action [action] [options]

Options:

• --list: List all supported actions
• --wait <seconds>: Wait time after action (default: 3.0)
• --monitor: Enable state monitoring output

Supported Actions:

• StandUp, StandDown, RecoveryStand
• Sit, RiseSit
• Hello, Stretch, Content, Scrape, Heart
• Dance1, Dance2
• FrontFlip, LeftFlip, BackFlip
• FrontJump, FrontPounce, Handstand

• examples/audio/internet_radio/stream_radio.py - Stream internet radio to the robot
• examples/audio/live_audio/live_recv_audio.py - Play robot audio live through host speakers
• examples/audio/save_audio/save_audio_to_file.py - Record robot audio to a WAV file

• examples/video/camera_stream/display_video_channel.py - Display live video frames with OpenCV

• examples/data_channel/move_test.py - Minimal Move command tester
• examples/data_channel/lowstate/lowstate.py - Comprehensive low-level state monitoring
• examples/data_channel/sportmodestate/sportmodestate.py - Monitor sport mode state values
• examples/data_channel/vui/vui.py - Control LED brightness, color, and flashing via VUI APIs

• examples/data_channel/lidar/lidar_stream.py - Basic LIDAR subscription and data printing
• examples/data_channel/lidar/lidar_performance_test.py - Measure LIDAR decoding performance
• examples/data_channel/lidar/plot_lidar_stream.py - Web-based LIDAR visualization via Flask/Socket.IO/Three.js
• examples/data_channel/lidar/record_lidar_pcd.py - Record LIDAR point clouds to PCD files
• examples/data_channel/lidar/view_pcd.py - View recorded PCD files

• examples/data_channel/robot_odometry/robot_odometry.py - Display robot odometry: pose and velocities
• examples/data_channel/robot_odometry/analyze_timestamp_drift.py - Analyze timestamp drift
• examples/data_channel/robot_odometry/timestamp_drift.py - Timestamp drift analysis tool

Control the Go2 robot using a USB/Bluetooth gamepad:

• apps/gamepad/gamepad_controller.py - Main gamepad control application
• apps/gamepad/gamepad_visualizer.py - Visualize and discover gamepad button/axis mappings
• apps/gamepad/gamepad_config.py - Gamepad configuration schema
• apps/gamepad/gamepad_mapping.yaml - Gamepad button/axis mapping configuration

Features:

• Joystick axes for continuous motion (forward/back, sidestep, yaw)
• Buttons and D-pad for discrete actions
• Configurable via YAML mapping file
• Optional obstacle-avoidance toggle

Control the Go2 using hand gestures captured from your webcam:

• apps/gesture/hand_gestures.py - Hand gesture recognition and robot control

Features:

• MediaPipe-based hand tracking
• Motion-to-action mapping:
  • Push hand forward → move back
  • Pull hand backward → move forward
  • Push hand down → StandDown
  • Push hand up → StandUp
  • Swipe left/right → side step
• Simulation mode (no robot)
• Two-hand requirement option to reduce false triggers

Combined video and LIDAR visualization using Rerun:

• apps/rerun/rerun_video_lidar_stream.py - Real-time video and LIDAR point cloud visualization

Features:

• Real-time video stream from robot camera
• Real-time 3D LIDAR point cloud visualization
• CSV read/write support for offline analysis
• Y-value filtering for LIDAR points
• Modular design (can disable video or LIDAR independently)

• Python 3.8+
• unitree_webrtc_connect (automatically installed as a dependency)
• Go2 robot with WebRTC enabled

For examples and apps, install optional dependencies:

# For examples with visualization
pip install ".[examples]"

# For apps (gamepad, gesture, rerun)
pip install ".[apps]"

# For development
pip install ".[dev]"

Set the robot IP address as an environment variable:

export ROBOT_IP="192.168.8.181"

Then run any example or app:

python examples/data_channel/move_test.py
python apps/gamepad/gamepad_controller.py

MIT

Based on the work from legion1581/unitree_webrtc_connect.