For the last while, almost everything in my business has been run by a single AI system I built on my own laptop. It writes and ships my brand's product pages, drafts and schedules content across channels, triages three inboxes, files my invoices, plans my week, and even maintains its own memory so it does not forget what we decided last month. It does most of this on a schedule, while I am asleep or at the bench. I finally sat down and wrote it up properly, as a research paper. Not because I need an academic stamp, but because writing it the rigorous way forced me to be honest about what actually works, what the numbers really are, and where it falls short. Here is the short version of what is in it. The idea. You do not need a heavy software framework to run a capable AI system. You need a well-organised set of plain text files. Folders and markdown files carry the instructions; the AI reads the right file at the right moment. The whole thing is something you could open and edit in a text editor. That is the same philosophy I teach inside MakerOS: your AI should be something you own and can read, not a black box you rent. What the system actually does, by the numbers. It carries 94 distinct skills and 15 specialised helper agents, all reached through one router. Over the measured period it logged more than five thousand scheduled runs across forty-one days. It keeps itself reliable (a 98 per cent clean rate across its skills), it keeps its own running costs in check with a budget governor, and it keeps its memory small enough to stay fast by archiving what it no longer needs, automatically. When it once let its own to-do file balloon, it noticed, and now trims it every night without me. Why I am sharing it with you, specifically. Two reasons. First, proof. When I tell you the AI methods in this community are real and not hype, this is the receipts: a working system, measured honestly, limitations and all. Second, this paper is essentially the blueprint behind MakerOS, the system you are learning to build for your own studio. You will recognise the ideas: own your tools, keep them readable, let the machine do the boring upkeep, and never let it touch anything irreversible without your say-so.

The commands, patterns, and workflows that get you from sketch to CAD-ready file.

Principles and Standards for Modern Jewellery Design and Production