Serving local AI models over my home network Draft
The goal
One machine in the house runs the models; every other machine — laptop, iMac, phone — just talks to it. No per-seat cloud subscription, no API meter running while I work.
Hardware and cost
- Mac Studio with M2 Ultra and 64GB of unified memory
Exact purchase details and current setup notes pending fact-check.
Software stack
- LM Studio — serves models over the local network through an OpenAI-compatible API, so anything that speaks the OpenAI API format can point at it
- Ollama — lightweight model runner, part of the regular rotation
What broke and how it was fixed
The real war stories go here after the author's fact-check — what failed on the first try, and what the fix was.
Verdict
Pending fact-check: who should copy this setup, and who shouldn't.