Run AI on your own hardware.

Real builds, real benchmarks, no cloud bill.

Latest builds Drafts

Every post below is a working draft, pending the author's fact-check.

Serving local AI models over my home networkDraft

A Mac Studio M2 Ultra with 64GB of unified memory, serving open-source models to every machine in the house through an OpenAI-compatible API.

Turning a 2017 iMac into a 24/7 homelab serverDraft

One aging all-in-one, Arch Linux, and a reverse proxy — the always-on box that runs the whole operation.

My local AI stack, and why I run my own infrastructureDraft

Ollama, Open WebUI, Stable Diffusion — the tools on the shelf, and the real reason the cloud doesn't get my work.