Back
~5 min read 9 Apr 2026

Claude + Obsidian have to be illegal

AI Summary

Combining Claude Code with an Obsidian vault creates a second brain that maintains itself automatically. By pointing Claude Code at your vault folder, it ingests articles, cross-references notes, runs weekly linting, and delivers morning briefings without manual upkeep. The key insight is that Claude handles all the maintenance work (summarizing, cross-referencing, filing) that causes most personal knowledge management systems to collapse over time.

Key takeaways

1

Most second brain systems fail because maintenance piles up. Claude eliminates that by making reorganization, cross-referencing, and filing a single command rather than manual work.

2

You do not need to open Obsidian at all. Point Claude Code at the vault folder, work from the terminal, and the knowledge base builds itself in the background.

3

Ingesting one article triggers Claude to link 10-15 wiki pages, surface unexpected connections, flag contradictions, and log what changed, so insights compound rather than disappear into chat history.

Original post

Claude + Obsidian should be illegal.

Let me tell you what my setup does before I explain how to build it.

Every morning I open my laptop, Before I type a single word, Claude already knows who I am, what I'm working on, every crypto tool I use, every open task, every article I've ever written, and every idea I've ever captured.

That's not a chatbot, that's a second brain that never sleeps, never forgets, and gets smarter every single day you use it.

Video thumbnail
This as an example what you can make in 5 minutes

Here's exactly how it works, and how you can build it yourself this week, it will take you few hours to save you THOUSANDS of hours:


How to set up your second brain in 5 minutes:


First let's download Obsidian (If you don't have it yet), few clicks and it's installed easily:

Article image
https://obsidian.md/

Now you need to create your second brain, that's called vault, you can name me however you want, mine is Leo's workspace:

Article image

Download Claude code for your desktop if you still don't have it:

Article image
https://claude.com/product/claude-code

Then go to Claude code and choose your vault:

Video thumbnail

And the most important step is Andrej Karpathyโ€™s prompt, which I split into two parts.

Copy the first part, then the second, and then put them together:

This section contains content that couldn't be loaded. View on X

This section contains content that couldn't be loaded. View on X

Now you need to help your second brain get info, nobody tells you this part:

You don't even need to open Obsidian, point Claude Code at the folder, work from the terminal, and your second brain quietly builds itself in the background.

Obsidian is just the window you look through when you want to see what's inside.

What actually goes in a second brain?

Think about everything you consumed in the last year that just disappeared:

> The book you finished and forgot.

> The podcast that changed how you think about something.

> Articles saved at 11pm that you never reopened.

> YouTube rabbit holes that taught you more than any course.

> Kindle highlights you marked and never looked at again.

> Research you did before a big decision.

> Old project notes.

> Lessons from things that went wrong.

All of that is sitting somewhere doing nothing. It belongs in your vault.

No existing material? Open Claude chat and talk for twenty minutes, your work, your goals, what you're building, what you're figuring out.

Save that conversation as your Memory file, that's enough to make your first session feel like Claude actually knows you.

The vault doesn't need to be complete to be useful. It just needs to be real.


The operations: what you actually do day to day

Ingest, you clip an article with the Web Clipper, it lands in your raw sources folder, you tell Claude:

This section contains content that couldn't be loaded. View on X

One article: Claude links 10-15 wiki pages, surfaces unexpected connections, flags contradictions, and logs exactly what changed.

Query: Ask the wiki, claude scans the index, pulls the right pages, and answers with citations.

Then it saves the best outputs back into the wiki, comparisons, analyses, new connections, so insights donโ€™t vanish in chat and your knowledge base compounds.

Lint: Once a week, run this:

This section contains content that couldn't be loaded. View on X

Your knowledge base stays healthy automatically, maintenance stops being your job.

The morning briefing set it once, runs forever:

This section contains content that couldn't be loaded. View on X

You set this up once, every morning it runs without you touching anything.

Process a call transcript and update your entire system:

This section contains content that couldn't be loaded. View on X

Every decision filed, every action tracked, nothing lost to chat history ever again


Why almost nobody has this


The reason this isn't everywhere yet is simple: the people who've built it aren't explaining it clearly, and the people who need it don't know it exists.

Most second brain projects die the same death. You start organized. Maintenance piles up, updating tags, keeping cross-references current, reorganizing when the structure evolves.

That's extra work on top of a full workload -> you skip it, the system degrades -> u go back to scattered notes. Six months later you try to rebuild it and the cycle repeats.

Claude breaks that cycle permanently, maintenance is just a command. Reorganizing your entire vault is a prompt, migrating from Notion?

One command processes every exported file, adds the right properties, and restructures everything into your new system.

The human's job is to curate sources, ask good questions, and think about what it all means. Claude's job is everything else, the summarizing, cross-referencing, filing, and bookkeeping that makes a knowledge base actually useful over time.

Vannevar Bush described something like this in 1945, a personal, curated knowledge store where the connections between documents are as valuable as the documents themselves.

He called it the Memex, the part he couldn't solve was who does the maintenance.

Now you know who does it: Build this once -> Use it forever, and it gets better every single day you add to it.

That's why it should be illegal.

- Leo

Join builders who get the best curated content in their inbox weekly.