01Private RAGPrivate RAG02Local modelsLocal models03Offline sessionsOffline sessions04Token costsToken costs05Turning it offTurning it off06Checking privacyChecking privacy
The hybrid approach

Local first. Cloud when it earns its keep.

Six chapters in two parts. The first four build on each other and set you up with a private, working AI on your own computer. The last two are about living alongside the AI you didn't choose. If you only read one, make it Chapter 02 — everything in Part One gets easier once you have a private model running on your own machine.


How to read these guides

Each chapter is written to be read in one sitting, at a kitchen table, with no assumed background. Technical terms are explained the first time they appear, and marked with a small tag like this:

term explained

Where a guide suggests trying something on your own computer, it says so clearly in a highlighted box — nothing is sprung on you.

The philosophy in one paragraph

Most of what people want from AI day-to-day — organizing thoughts, drafting a letter, searching their own notes, thinking through a decision — can be done entirely on a personal computer, for free, with total privacy. Save the large, cloud-based "frontier" models for the handful of tasks that truly need extra horsepower, and use them on purpose, not by default.