The short version: minibases are small, focused knowledge bases your LLM can actually read. Here's what they are, why this blog exists, and who's behind it.
What is a minibase
A minibase is a small, curated collection of notes on one topic, built so a language model can read it and reason over it. Not a giant company wiki. Not every PDF you've ever clipped. Twenty to fifty files of carefully chosen, plainly written markdown, small enough to summarize from memory, sharp enough to make a model think like a friend who's done the homework.
Why this blog
LLMs are only as smart as the context you give them. Most people give them the chat window. The few who don't, quietly outperform everyone else. This blog is the playbook. How to pick a tool. How to capture from the web. How to structure notes for retrieval. How to plug the result into a real model. One post per question. No frameworks, no twelve-step systems, no acronyms.
Who built it
I built minibase on my weekends. I write here as @jswallez. I also built Save, a browser extension that turns any web page into clean markdown, basically the tool I use to feed my own minibases. If a post is useful, tell me on X.