Skip to content

Open Source AI Robot Software for InMoov | LLM + Computer Vision + Voice Interaction | Python | Radxa | NVIDIA Jetson | Raspberry Pi

Notifications You must be signed in to change notification settings

alexbuildstech/nova

Repository files navigation

Nova - Open Source AI Robot Software for InMoov

Python License Status Platform

Nova is a state-of-the-art, independent software stack designed specifically for the InMoov open-source 3D printed robot platform. It transforms static animatronics into fully autonomous, conversational humanoid robots using advanced Artificial Intelligence.

Unlike traditional script-based robotics, Nova leverages Large Language Models (LLMs) like OpenAI GPT-OSS-20B and Computer Vision models like Google Gemini 2.0 Flash Vision to enable real-time, dynamic interaction. Whether you are running on a high-end PC, an NVIDIA Jetson, or a powerful SBC like the Radxa rock 5C, Nova provides the intelligence your robot needs.

🚀 Nova is the next-generation independent AI stack for InMoov, optimized for modern LLMs and cloud-native AI services. Built from the ground up for flexibility and performance.


🤝 Sponsors & Partners

This project is proudly supported by industry leaders in making and robotics:

Robotics & Open-Source Hardware Advanced 3D Printing Materials High-Performance SBCs

🔧 Key Hardware from DFRobot

Component Product Use Case Link
Main Controller DFRduino Mega2560 (×2) Servo control & sensor processing View Product →
Gripper Servos DSS-P05 Standard Servo 5kg (×2) Precise gripper finger movement View Product →
Force Sensors Circular Force Sensor 7.6mm (×2) Tactile feedback for grip pressure View Product →
Stereo Vision USB Camera 720p Wide-angle (×2) Depth perception & object recognition View Product →

💡 Recommended DFRobot Products

For builders looking to expand with arm articulation:

Component Product Torque Best For Link
Arm Joints High Torque Waterproof Servo 20kg 20 kg·cm Shoulder & wrist rotation View Product →
Bicep/Elbow High Torque Servo with Encoder 25kg 25 kg·cm Arm lifting & elbow bend View Product →

💬 "DFRobot's ecosystem made building Nova's manipulation system possible. Their servo lineup and force sensors are perfect for humanoid robotics."


📑 Table of Contents


🚀 Key Features

  • 🧠 Advanced LLM Integration: Powered by OpenAI GPT-OSS-20B (via Groq) for ultra-low latency, human-like conversation.
  • 👁️ Computer Vision & Perception: Uses Google Gemini 2.0 Flash Vision to analyze the visual world, describe objects, and read text in real-time.
  • 🌐 Real-Time Internet Access: Integrated Google Search allows Nova to answer questions about current events, weather, and news.
  • 🤖 Autonomous Face Tracking: PID-controlled servo tracking ensures Nova maintains eye contact, creating a "living" presence.
  • 🗣️ Natural Voice Interaction: Features high-fidelity Speech-to-Text (STT) and Text-to-Speech (TTS) for seamless verbal communication.
  • 💻 Web Control Dashboard: A local web interface for easy monitoring, debugging, and control.

🛠️ Hardware Requirements

Nova is optimized for the InMoov robot head but is compatible with various animatronic setups.

  • Compute:
    • Recommended: Linux/Windows PC with NVIDIA GPU.
    • SBCs: Compatible with Radxa rock 5C, Raspberry Pi 5, or NVIDIA Jetson Orin Nano.
  • Microcontroller: Arduino Uno or Mega (for servo control via Serial).
  • Vision: USB Webcam (Logitech C920 recommended).
  • Audio: USB Microphone & Speakers.
  • Mechanics: Standard Hobby Servos (MG996R/SG90) for InMoov neck and eye mechanisms.

📦 Installation

  1. Clone the Repository

    git clone https://github.com/alexbuildstech/nova.git
    cd nova
  2. Install Python Dependencies

    pip install -r requirements.txt
  3. System Setup Ensure you have ffmpeg installed for audio processing:

    sudo apt install ffmpeg  # Linux
    # or download for Windows

⚙️ Configuration

Nova uses a central configuration file for security and ease of use.

  1. Open config.py.
  2. API Keys: Enter your Groq API Key and Google Gemini API Key.
  3. Hardware: Set your CAMERA_INDEX and SERIAL_PORT.
# config.py example
GROQ_API_KEY = "your_key_here"
CAMERA_INDEX = 0  # 0 for default webcam
SERIAL_PORT = "/dev/ttyACM0"  # or 'COM3' on Windows

🔮 Roadmap

  • MyRobotLab Integration: Full plugin support for the MyRobotLab ecosystem.
  • Gesture Control: Hand tracking and arm movement synchronization for InMoov arms.
  • Local LLM Support: Offline inference using Llama 3 via Ollama.
  • Emotion Engine: Dynamic facial expression mapping based on sentiment analysis.

📄 License

This project is open-source and available under the MIT License.


Keywords: InMoov, AI Robot, Python, LLM, Computer Vision, OpenAI, Gemini, Robotics, Animatronics, Open Source, Artificial Intelligence, Face Tracking, Voice Assistant, Radxa, Polymaker, DFRobot, MyRobotLab, Raspberry Pi, NVIDIA Jetson.

Releases

No releases published

Packages

No packages published

Contributors 2

  •  
  •