About Hex Index

Hex Index is a reading library for people who want to go deep. Long-form articles and video transcripts from independent writers and educators — curated, enriched with Wikipedia deep dives, and adapted for an enjoyable reading experience. No ads. No tracking. No algorithmic feed. Just good writing, ten minutes or longer, from sources worth your attention.

Why This Exists

I love Substack and YouTube. I subscribe to dozens of independent writers and creators — people who do real reporting, real analysis, real teaching. I can spot a quality educator. I know what careful thinking looks like in print.

But the reading experience frustrated me. Substack's web interface is a single-page app, which means I have to Command-click every article to open it in a new tab if I want my text-to-speech extension to work. Then I'm managing twenty tabs and forgetting which ones I've read. The subscription feed shows everything — audio posts, short updates, paywalled teasers with three paragraphs of free content. There's no way to filter for what I actually want: text articles with ten minutes or more of free reading material.

YouTube is worse. Brilliant long-form content buried under an interface designed for short attention spans. No way to read a transcript as an article. No way to collect video essays alongside written ones.

I wanted Substack and YouTube in one place, from a curated list of sources, filtered to the good stuff. So I built it.

The Deeper Problem

There's something else. I'm worried about the state of journalism. Monopoly ownership. Fascist attacks on the press. Both sides parroting party lines instead of investigating. The constant churn of headline news — one crisis to the next with no context, no history, no depth. Everything is fed to you. Everything is brief. Everything is designed to be consumed and forgotten.

Hex Index is my small answer to that. A spotlight on independent, top-notch coverage. Slow burn. Context. Thoughtful. Long form. The kind of reading that The Week magazine used to offer in its heyday — a curated window into the best thinking happening right now.

How It Works

Every hour, the system pulls new articles from RSS feeds and YouTube transcripts. It filters for text content with ten minutes or more of free reading — paywalled articles are fine as long as the free portion is substantial. For each article, it identifies three related Wikipedia topics, scrapes them, and rewrites the encyclopedic content as engaging essays — the kind of background reading that makes a good article great. Every illustration is generated specifically for its article.

Articles are also editorially adapted: restructured for clarity, converted to third person, and enhanced with light counterpoints and editorial judgment. The original is always linked. We're curators and editors, not replacements.

The entire system runs on a Mac Studio with a local language model. The public site is static HTML — no database, no server, no tracking, no cookies. Just words.

About Brian

Brian Mabry Edwards writes about human-AI collaboration at LLuMinate, where he coined the term to describe "the act of writing with the assistance of large language models — the spark of human intention meeting the strange mirror of machine language."

He also makes music at jalopy.music.

Happy reading.