Local AI with Gemma by Google

This project aims to make Liquid Galaxy easier to use by building a system that can automate content creation and control the screens using simple natural language commands. Currently, users need technical knowledge to create KML files, manage content, and operate the system, which makes it difficult for beginners. To solve this, I will build an AI server that runs on a Single Board Computer and acts as an assistant for Liquid Galaxy. Users will be able to give commands using chat or voice, and the system will handle tasks like generating KML visualizations, sending content to the screens, and controlling the setup. The system will be modular, so different features (like visualization, control, and communication) can be added or improved easily. It will also support both local processing and external services depending on what works best with the available hardware. This server will live on the same LAN as the liquid galaxy rig. The main deliverables include: A working AI agent server running on a Single Board Computer Chat and voice interfaces for user interaction, external services like telegram, discord, etc will also be supported. A media pipeline to generate and send KML, images, and other content to Liquid Galaxy A content builder for creating simple apps, and interactive outputs Real-time data visualization from open data sources (aircraft, satellites, disasters, etc.) Automated storytelling and guided tours based on user input (eg- historical events) News and event-based visualization on the globe A knowledge assistant using Liquid Galaxy documentation An interactive troubleshooting system for diagnosing and fixing issues User friendly instructions via .md files, to help new developers get started Agent personality customization Complete documentation, deployment scripts, and setup guides The final result will be a simple, flexible, and easy to use system that helps users and contributors interact with Liquid Galaxy.