Deep dives
The main pages explain the method without taking anything for granted. Here is what sits underneath, for anyone who wants to see it: the exact shape of the files, the rules the program applies, and the reasons behind the choices. You don't need to read it to use IllAIra.
Where to upload, model by model
Gemini Notebook, Claude, ChatGPT, Copilot, Gemini, Grok: where the files go, and the line to write so they aren't treated as attachments.
It isn't a Skill
Claude Skills, custom GPTs, Gems: they do a different job. Pros, cons, where they coexist — and the objections, one by one.
It isn't the graph you've already seen
How the Sferografia differs from a graph of notes or a personal wiki — and why the difference isn't cosmetic.
Structure is enough
A fruit fly's brain switched on inside a computer: what it teaches about making a mind exist on a substrate that isn't its own — and what it doesn't.
The permission zones
Five levels, from «only you may touch it» to «the AI writes whatever it likes in here». What actually changes between one and the next, and how you choose.
Anatomy of a file
How a Reminder is made, line by line: the first line that says what kind of file it is, the cores, the modules, the logs. And what happens when you get it wrong.
Missing something you needed? Tell us: these pages come from the questions that arrive, not from a table of contents decided in advance.