Studio Sentinel

Studio Sentinel turns your iPhone into a real-time monitoring dashboard for your Mac Mini or Mac Studio running local AI workloads. If you run Ollama, LM Studio, llama.cpp, or any other local LLM server on an Apple Silicon Mac, you know the challenge: you start a long inference job, walk away, and have no idea what is happening. Is the model still running? Is the machine thermal-throttling? Did memory pressure kill the process? Studio Sentinel answers all of these questions from your iPhone - no cloud, no subscription required for core features, and no data ever leaves your local network. WHAT IT MONITORS CPU - Real-time usage across all performance and efficiency cores. See exactly how much of your M2 Ultra or M3 Max is being consumed by inference workloads versus background tasks. Memory - Unified memory usage, pressure level (nominal / warning / critical), wired pages, and free memory. Essential for knowing whether your machine has headroom for another model. GPU - Apple Silicon GPU utilization estimate and unified memory allocation. Understand how hard your Neural Engine and GPU cores are working during inference. Thermal - Live thermal state from Apple's kernel. Know immediately if your Mac is throttling due to heat before it impacts your inference speed. Disk and Network - Storage capacity at a glance and real-time network throughput on your primary interface. AI WORKLOAD INTELLIGENCE Ollama - Live status of your Ollama server including API availability, response latency, models currently loaded in unified memory, and all models available on disk. Docker - Running containers with health status and port mappings. See your Open WebUI, Qdrant, ChromaDB, and other AI containers at a glance. Process Monitor - Full process table sorted by CPU or memory. Every AI-related process is automatically identified and highlighted. Health Score - A single 0-100 score combining thermal state, memory pressure, Ollama availability, and Docker status into an at-a-glance inference readiness indicator. HISTORICAL CHARTS (PRO) Visualize CPU, memory, and GPU trends over time. Identify inference bursts, memory leaks, and thermal patterns with interactive charts. ALERTS (PRO) Set thresholds for memory pressure, CPU usage, thermal state, and health score. Get notified the moment your Mac needs attention. HOW IT WORKS Studio Sentinel Agent runs as a lightweight menu bar app on your Mac. It exposes a local REST API over your LAN. The iOS app discovers it automatically via Bonjour - no IP addresses, no configuration. Just install both apps on the same network and they find each other. All communication is secured with a one-time pairing token stored in the iOS Keychain. No data leaves your local network. No cloud servers. No accounts required. DEMO MODE No Studio Sentinel Agent nearby? Tap "Try Demo" on the device list screen to explore a fully simulated Mac Studio M2 Ultra with Ollama running llama3.1:70b and codellama:13b. Every screen is populated with realistic data so you can evaluate the app before connecting a real machine. REQUIREMENTS iPhone or iPad running iOS 17 or later Studio Sentinel Agent installed on a Mac running macOS 14 (Sonoma) or later Both devices on the same local network Studio Sentinel Agent is a free download available at github.com/govmed/MacMiniSystems