I built Aegis AI – An Agentic Home Security w/ GPT+Local VLM on Mac/PC
Local VLM + GPT agentic reasoning replaces dozens of motion alerts and scattered camera apps.

Local VLM + GPT agentic reasoning replaces dozens of motion alerts and scattered camera apps.
Homeowners wanting intelligent, unified security monitoring without subscription bloat
Frigate (open-source NVR) · Ring Protect · Nest Aware
Aegis AI is a desktop app that turns your existing cameras into an intelligent security system — the world's first AI Security Agent for your home. It runs on your Mac or PC (even a Mac M1 Mini with 8GB RAM can deploy a LFM2.5 model for video analysis), connects to any camera you already own — and even your laptop's webcam or an old iPhone can become a security camera. Your agent stays connected to your phone through Discord, Telegram, or Slack, so you're always in the loop.
The World's First AI Security Agent for Your Home Home security today means dozens of useless motion alerts a day, footage scattered across multiple apps with multiple subscriptions, and no way to just ask "did anyone come to the door today?" Aegis changes that. It's an AI agent that watches, understands, and talks to you — not a dumb motion detector.
How Aegis Works Connect any camera — Ring, Blink, DaHua, HikVision, Reolink, any RTSP/ONVIF IP camera, your phone, your webcam, even a retired iPhone or iPad sitting in a drawer. Aegis unifies them all into one place.
AI watches and understands — Choose how you want your AI to run: locally via llama-server with vision models from HuggingFace — SmolVLM2, LLaVA, MiniCPM-V, Qwen-VL, LFM, and more — or through GPT Vision / Google APIs with your own key. Aegis doesn't just flag motion — it tells you who's there and what's happening.
Smart alerts, not spam — Instead of 47 notifications about wind and shadows, you get meaningful ones: "UPS driver at the front door" or "your kid just got home from school." Behind this is an agentic framework with an advanced memory and knowledge system — Aegis learns who's family, what's routine, and what's actually unusual. It deduplicates events, builds context over time, and makes security decisions, not just detections. Alerts go to Slack, Discord, or Telegram — wherever your family already is.
Ask it anything — Missed something? Just ask. "What happened at the front door today?" Your agent watched everything and gives you a real answer.
Everything stored locally — Every clip saved on your machine. Instant playback, no buffering, no subscription. One unified timeline across all cameras — scrub through your day and find the exact moment.
Why I Built This I have a mix of Ring, Blink, and IP cameras. I was tired of useless alerts, footage locked behind subscriptions, and having no way to just ask what happened while I was away. So I built the agent I wanted.
Current Status The app is functional in beta release and I use it daily. Multi-camera integration, local and cloud AI, unified timeline, chat interface, smart alerts, and cross-platform support (Mac, Windows, Linux) are all working. Currently adding more camera integrations and a mobile companion app.
Try it: https://www.sharpai.org
Happy to answer any questions!
Local VLM + GPT agentic reasoning replaces dozens of motion alerts and scattered camera apps.
Replaces Google Nest's $20/month cloud analysis with local Qwen 35B and a 3D-printed head.
Reproducible builds across the entire stack—rare for consumer IoT security.
Reproducible builds across entire stack with E2E encryption, unlike Ring or Nest.
One-call auto_instrument() beats manual Guardrails AI integration.
Security-by-construction language for AI agents with Z3 contract verification.