Video Library
Experiments • Prototypes • Demonstrations
Explore our collection of research videos documenting AI experiments, hardware prototypes, and breakthrough demonstrations from theLAB.
Talking to My AI Feels Like Talking to a Person Now
Realtime mode for ADI — continuous vision, conversational pace, no tool calls, no "send message" button. Just a live conversation with something that's actually watching. Tracks intent through speech (knew what I meant when I pointed at my hat and announced I was taking it off), recognizes detailed objects (identified the cat statue as Bastet, Egyptian goddess depicted as a lioness or cat), and reads OCR-grade text (picked $7.98 off a Lovecraft paperback price tag). All running on theLAB hardware.
ADI — Fully Local Voice Assistant with MCP Servers
ADI (Advanced Data Intelligence) is a fully local voice assistant built at theLAB with OpenWakeWord, Parakeet, Ollama, OmniVoice, and ComfyUI + FLUX. The demo shows MCP servers for web search, knowledge graph memory, file operations, shopping lists, tasks, and multimodal vision-to-image generation running on local hardware.
Vision-TTS-Art — AI That Sees, Speaks, and Creates
A multimodal AI system that combines computer vision, text-to-speech synthesis, and generative art capabilities. Watch as AI sees the world, describes it aloud, and creates visual art based on its understanding.
ADI OCR — Local AI Document & Image Analysis
Local AI-powered document and image analysis using MiniCPM-V. Extract text, analyze documents, and process images entirely on your own hardware with no cloud dependencies.
ADI — Advanced Data Intelligence
Advanced Data Intelligence system from theLAB. A comprehensive AI platform for intelligent data processing, analysis, and insights generation.
AI Network Analysis
AI-powered network analysis and monitoring. Leverage machine learning to understand network traffic patterns, detect anomalies, and optimize infrastructure performance.
LLaVA Vision AI — Real-Time Local Computer Vision
Real-time local computer vision system powered by LLaVA. Process and understand visual input instantly on your own hardware without cloud latency or privacy concerns.
AI Hallucination Checker — Extended Demo
Extended demonstration of the AI Hallucination Checker. Validate AI-generated content for accuracy and detect potential hallucinations before they cause problems.
MOBILE.NODE — Llama 3.2 + Whisper + CoquiTTS on Tablet
MOBILE.NODE demonstration running Llama 3.2 1B, Whisper speech recognition, and CoquiTTS text-to-speech entirely on a tablet. Full AI stack on portable hardware.
VADER — Cyber Security Research Series
VADER (Vulnerability Assessment Detection Enhanced Reasoning) is an AI-powered network vulnerability scanner designed to detect, analyze, and report potential security threats in real-time. Educational series combining intelligent reasoning with advanced detection.
SABER — Software-Assisted Broadcast Enhanced Radar
SABER is a cutting-edge system designed to enhance radar capabilities using advanced software integration. Combines traditional radar technology with real-time data processing, visualization, and AI-powered analytics for improved situational awareness.
GibberLink — AI Sound-Based Communication
Two local AI systems communicate using GGWave, a protocol that transmits data through sound frequencies. Powered by GibberLink, devices exchange messages over ultrasonic or audible frequencies — no Wi-Fi, Bluetooth, or internet required.
vendOS — Raspberry Pi Point of Sale System
vendOS is a sleek, self-serve Point of Sale kiosk system designed for office or small-space vending. Powered by Raspberry Pi, it combines barcode scanning, interactive touchscreen interface, and optional PayPal QR code payments on an honor system framework.