Make Your Second Brain Portable: Local First, Never Synced

How do you make a second brain portable? Keep the real one local, as plain files in folders on your own computer. Then send copies out on purpose: a private backup, a cleaned copy for your team and agents that write what they learn back home. What you never do is hand the original to an app that syncs it for you.

That's how mine works. As I told a group I trained this fall, "one of the things that I knew I wanted to make 100% sure is whatever I was creating is portable."

It Started Local, and It Stays Local

My brain has three versions online. "But it all started local and I always keep it local."

That's the whole foundation. If you've read What Is a Second Brain?, you know mine is a set of folders full of plain files. Plain files go anywhere. Any computer can open them, any AI can read them and no company can lock them away from you.

Everything else in this post is a copy of that one folder.

Copy One: A Private Backup

The first online version is a private GitHub repository. GitHub is a service built to store versions of files, and private means only I can see it.

I back the brain up to it "every single day or almost after every project that I do." So it keeps adding context, and it keeps saving.

That gives me two wins. If this computer ever dies, the brain doesn't. And I can open the same repo on my laptop, or any computer, and pick up where I left off.

Copy Two: A Cleaned Copy for the Team

The second version is a Google shared drive. It's "a mapping of the main brain, minus anything that is a security risk or personal information like credit cards, social security, any of that."

That drive connects to Claude, because my team works in Claude Team projects. "And so when anybody works in a team project, it looks at the Google main brain to pull information into, now it's for the team instead of just for me."

Read that list of what stays out one more time. A team copy isn't a full copy. You decide what the team never needs to see before anything leaves your computer.

Copy Three: Agents That Write Back Home

The third version is my agent team. I layered Grokbot on top of my main brain. Then I had each agent build its own Karpathy-style wiki knowledge base, one per client or one per tool it was going to use, "so they could get super smart on the things that they needed to know."

If the intake, raw and wiki pattern is new to you, Intake, Raw, Wiki walks through it. And Agents, Skills and Grokbot shows how the two teams share one brain.

Here's the part that keeps it portable. Anything the agents learn goes back to the main brain. "So if all of a sudden I left Grokbot, remember portable, if I left Grokbot, everything that we've worked together would at least still be in that main brain, I could pick the new thing and start running."

That's the test I'd hand you. If you walked away from your favorite AI tool tomorrow, would everything you built together still be yours?

Portable for Model, Too

There's a second kind of portable. Not where the brain lives, but which AI reads it.

"I can log in with my ChatGPT account and work with my main brain. I can log in with my Anthropic account and work with my main brain."

Because the brain is plain files, no single model owns it. I can bring a different AI to the same brain, and it knows what the last one knew.

The OneDrive Mistake

I didn't learn this the easy way.

"This is when I was a youngin', you know, a year ago, year and a half ago. I thought it'd be real smart and build the first version of Echo on Microsoft's OneDrive." The plan was to auto sync, so I could sit at one computer or another and use the same Echo.

It went wrong. When another human in that training described cloud sync trouble, I told them, "I ran into the same situation."

So here's my rule now: "I don't sync it automatically with Google. I don't sync it automatically with OneDrive."

How the Copies Stay Current Without Syncing

If nothing syncs, how does the team's shared drive stay up to date? An agent does it, on a schedule.

The main brain sits in a folder that doesn't sync anywhere. Then I have agents, or a routine, or a schedule, that keep that separate Google shared drive current. In my words, it's a drive "that isn't synced with anything," and "the agent keeps it up to date with whatever we add."

There was a one-time porting to start it. After that, "nothing is being synced, nothing ever moves."

The difference matters. A sync app decides what moves and when. An agent on a schedule copies what I've said it can copy, to the place I've said it goes, and the original never leaves home.

Look at where my second brain lives right now. Tell me whether any of it sits inside a folder that a cloud sync app like OneDrive, Google Drive, iCloud or Dropbox manages.

Then help me make it portable. Keep the original as plain files in a local folder that doesn't sync anywhere.

Set up a private backup I can update every day or after every project, and explain each step in plain English.

Then list what I should keep out of any copy my team or other tools can see, like passwords, financial details and personal information. Ask me one question at a time about anything you're unsure of, and show me a plan before you move or copy any file.

This is the starter version. The deeper conversation starters for building your brain live in the AI Second Brain Workshop and my coaching.

Who This Is Not For

If your notes live happily in one app and you never plan to switch tools, computers or models, you may not need any of this.

But if you're building something you want to keep for years, start local. It's a lot easier to begin with plain files than to dig your brain back out of somebody else's cloud later.

Frequently Asked Questions

What does it mean for a second brain to be portable? It means the brain isn't stuck in one place. Mine stays local as plain files, gets backed up online and can be read by more than one AI.

Should I keep my second brain in OneDrive or Google Drive? I don't. I built the first version of Echo on OneDrive to auto sync between computers and ran into trouble. Now I keep the original in a folder that doesn't sync anywhere.

How do I back up a second brain? I back mine up to a private GitHub repository every day or after almost every project.

How does my team use my second brain without seeing everything? I keep a separate shared drive that maps the main brain, minus anything that's a security risk or personal information. My team's Claude Team projects read that copy.

What happens to my work if I stop using an AI tool? If your agents write what they learn back to your main brain, it stays with you. If I left Grokbot, everything we built together would still be in my main brain.