The argument in one line.
The most durable second brain is one that maintains itself: drop raw files in, let an AI agent synthesize your wiki automatically, and your knowledge base compounds without the maintenance cost that kills every complex setup.
Read if. Skip if.
READ IF YOU ARE…
- You have abandoned an Obsidian or Notion second brain because the maintenance felt like a second job.
- You use Claude Code or Claude Cowork daily and want project context to persist and grow without manual documentation.
- You run a solo business or content operation and want a queryable knowledge base you can build in under an hour.
- You are curious about scheduled AI agents doing practical background work on local files.
SKIP IF…
- You already have a working second brain system and are not looking to switch.
- You need a team-wide, cloud-synced knowledge management platform — this demo is local-first.
The full version, fast.
Most second brain systems fail because they require constant manual curation. The SimpleBrain setup, built on Andrej Karpathy's self-documenting knowledge base idea, reduces that to one action: drop anything into a raw folder. From there, an AI agent reads every file, synthesizes clean notes into a wiki folder, and moves processed files to archive. A CLAUDE.md instruction file tells the agent how to behave; a translate.md prompt is the reusable trigger. Once a scheduled daily task is wired up in Claude Cowork, the system runs without the user doing anything beyond feeding it new raw material.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →Where the time goes.
00:00 – 01:02
01 · Why most second brain setups are too complicated
Hook frames the problem: everyone is building AI second brains with Obsidian, most setups are overcomplicated, host promises a simpler path.
01:02 – 03:34
02 · What Obsidian actually is
Demystifies Obsidian as just a markdown editor. Shows the knowledge graph visual and argues it is the end state, not the starting point.
03:34 – 06:39
03 · Introducing SimpleBrain
Introduces SimpleBrain as a minimal second brain built on Andrej Karpathy's self-documenting knowledge base idea. Three core concepts: raw dump, AI builds the brain, extend or query.
06:39 – 08:49
04 · The folder structure — five folders, three files
Walks through raw, wiki, archive, prompts, and projects folders plus CLAUDE.md, agents.md, README.md. Explains how each element functions in the loop.
08:49 – 16:26
05 · Setting up in Claude Cowork
Creates a new project in Claude Cowork, connects the SimpleBrain folder, tours the README, CLAUDE.md, agents.md, and translate.md files in detail.
16:26 – 20:50
06 · Live demo — raw to wiki
Feeds the Build Great Products website URL to Claude Cowork as a raw input, watches it generate a markdown file, then runs the translate prompt and watches wiki files appear.
20:50 – 23:08
07 · Setting up a daily scheduled automation
Creates a scheduled task in Claude Cowork that runs translate.md daily at 9AM — showing the full hands-off automation loop.
23:08 – 25:30
08 · Extension ideas
Four automation ideas: scheduled translation, voice memo transcription via Whisper, YouTube/podcast transcript ingestion via yt-dlp, weekly wiki synthesis summary.
25:30 – 26:24
09 · Where to get the files
Points to the SimpleBrain GitHub repo and the Skool community for download and support.
Lines worth screenshotting.
- Obsidian is just a markdown editor — there is nothing it can do that a plain folder of text files cannot.
- The maintenance cost of someone else's second brain template is higher than the setup cost, which is why most people abandon it.
- An AI agent that auto-documents your raw inputs removes the only part of personal knowledge management that people actually skip: the writing-it-down step.
- Five folders and three files is the entire structure — raw, wiki, archive, prompts, projects, plus CLAUDE.md, agents.md, README.md.
- The translate.md prompt is the whole engine: read raw, write wiki, move processed files to archive, commit to git.
- Git version control means nothing in the second brain is ever permanently lost — every state is recoverable.
- A daily scheduled task in Claude Cowork costs almost nothing to run and eliminates the forget-to-process-notes failure mode entirely.
- Voice memo transcription, YouTube transcript ingestion, and weekly synthesis summaries are all one prompt file away from being automated on top of the same base structure.
- You can build a separate SimpleBrain instance per project rather than one monolithic system, which makes context isolation trivial.
- The cloud version of this system — a team-accessible second brain synced via GitHub — is a startup idea the creator openly flags as wide open.
The second brain that runs itself
WHAT TO LEARN
The reason every Obsidian setup fails is the same reason journals fail — maintenance is the product, and AI eliminates it entirely.
01Why most second brain setups are too complicated
- Most second brain systems fail at the maintenance step, not the setup step — the tool is rarely the problem.
02What Obsidian actually is
- Obsidian is a markdown editor with link-between-files support — the elaborate knowledge graphs you see in demos are years of manual curation, not a starter template.
- Any tool that requires you to understand someone else's organizational system before you can use it has already failed you.
03Introducing SimpleBrain
- The self-documenting knowledge base idea is simple: dump raw content, let the AI write the wiki, extend on top of whatever it builds.
04The folder structure
- Five folders and three files is a complete second brain system; complexity beyond that is a feature request, not a prerequisite for getting started.
05Setting up in Claude Cowork
- CLAUDE.md and agents.md are instruction files that front-load all behavioral rules so you do not have to re-explain context in every prompt.
- The translate.md prompt has seven rules that prevent the most common failure modes: never delete from raw or archive, always read wiki before writing, commit after changes.
06Live demo — raw to wiki
- The only step people consistently skip in knowledge management is writing things down — an AI translate prompt automates exactly that step, turning raw dumps into structured wiki entries without user effort.
07Setting up a daily scheduled automation
- A scheduled daily agent costs almost nothing to run and eliminates the core failure mode of manually-maintained knowledge bases: the days when you forget to process your notes.
08Extension ideas
- Voice memos, YouTube transcripts, and podcast audio are all valid raw folder inputs — the same translate loop processes them automatically once the file lands in the folder.
- A weekly synthesis prompt that reads all wiki changes from the last seven days generates a standing digest of what your knowledge base learned that week.
Terms worth knowing.
- SimpleBrain
- A minimal second brain structure — five folders and three files — where an AI agent reads raw inputs and automatically generates a clean wiki, eliminating manual documentation.
- translate.md
- A reusable prompt file stored in the prompts folder that instructs an AI agent to convert everything in raw into organized wiki entries and move processed files to archive.
- CLAUDE.md / agents.md
- Instruction files that Claude or any AI agent reads before executing prompts; they define user context, allowed tasks, wiki voice rules, and out-of-scope actions.
- wiki folder
- The AI-maintained knowledge base within SimpleBrain — one markdown file per topic, merged not overwritten, updated automatically from raw inputs.
- Andrej Karpathy second brain post
- A widely shared piece by AI researcher Andrej Karpathy describing a self-improving, self-documenting local knowledge base — cited as the conceptual foundation for SimpleBrain.
- Claude Cowork
- Anthropic's agentic desktop interface for Claude that supports persistent projects, file context, and scheduled tasks — used in this video to demonstrate the SimpleBrain setup.
Things they pointed at.
03:34linkAndrej Karpathy viral second brain post
23:40toolOpenAI Whisper (audio transcription)
Lines you could clip.
05:06
“You don't have to maintain any of this system really, the AI is doing it for you.”
07:10
“Five folders and three files. That is literally it to start off with.”
22:11
“That is gonna run every day at 9AM inside of the simple brain project.”
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
metaphoranalogystory
00:00Everyone is talking about creating their own second brain using AI tools like Obsidian and Claude code. But in this video, I wanna break down why most second brain setups are just way too complicated and how to get set up with a second brain that actually works. Because it's super easy to set up someone else's second brain system and not have a clue about what's going on.
00:20So in this video, I wanna show you why most second brains don't actually work, the super simple second brain setup that you can get up and running with in minutes, and how to unlock the power of your second brain. And we're gonna do this without Obsidian and without all of the complicated setups that you see most other people doing.
00:39If you don't know me, my name is Chris and for the last fifteen years, I've been designing apps and advising startups on product and design. And I've helped thousands of people through my community learn how to build with AI, launch, and get real customers.
00:51And if you're interested in learning more about that, you can click the link in the description down below. Otherwise, let's jump straight into this second brain setup and get you up and running with a second brain that actually works in minutes. Now you've most likely seen this icon before, which is the icon for Obsidian, with everyone talking about creating their AI second brain using Obsidian and tools like ClaudeCode as well.
01:11And Obsidian, if you don't know what Obsidian is, you might be seeing all of these videos and thinking, well, I must have to use Obsidian. Obsidian must be the way that I set up my second brain. It's the second brain tool that people are using.
01:22But Obsidian is basically just a markdown editor, and all that means is that it edits text files in a markdown file format. That's all a markdown document is.
01:32It's just a text file with a bunch of text in it. Now there are a bunch of really great features with the tool like Obsidian. You can add extensions to do specific things and you can link between all your markdown files so you create this kind of crazy looking knowledge graph that looks a little bit like this.
01:48And you've probably seen demos and videos of all these crazy connections and things going on that people have set up in Obsidian. All these documents are linked to each other and all this crazy amount of knowledge stored in one place. And you might have even looked at it and thought, that looks amazing.
02:03How can I get that for myself? But I wanna tell you that if you wanna create a second brain, you don't need to start off by building something that's this complex. You don't need to start off by using a tool like Obsidian.
02:15You can just start off with a simple folder structure and a set of text markdown files that you can use to run your second brain and start to build up a knowledge base that you can reference with AI tools, that you can use to build automations, that you can use to document your entire business, and basically run your business from this second brain.
02:36Now the second brain setup might end up looking something like this, but to start off with, you absolutely don't need to do this. So it doesn't have to be like this. You don't need to start off with Obsidian.
02:47You don't need complicated setups. Don't need to try and understand someone else's brain either. Most people are pulling together packages that you can install for yourself that have all of these different folders and files and scripts and everything going on inside of them.
03:02You don't need to do that, and you don't need to be constantly fixing your second brain, which is something that can happen if you're using these other second brain setups that other people are talking about. And so let's talk about how you can actually set up your own second brain in a very, very simple way.
03:18I know I said second brain a lot in this video, and there are only a few key concepts to understand here. So if you're confused by all of these other second brain videos and you're thinking about how do you get started and create a second brain that actually works for you, we really just wanna focus on a few key concepts.
03:34So I created this thing here, which I have used for a bunch of different projects as well, which I call simple brain, which is basically a simple version of a second frame. Simple brain sounds a bit dumb, but I kinda like that in the name. And this is the simple second brain setup.
03:49And it's basically built on top of Andrei Kapathi's idea for this self improving, self documenting knowledge base in this post that went viral and got basically over 21,000,000 views.
04:03And in his model here, and a lot of people had talked about this model as well, he basically had a few key things. And the three main things that I would recommend starting out with in this simple second brain setup is basically a folder where you put all of your raw files, and you can put PDFs, markdown files, documents, CSV files, JSON files, website URLs, all of this stuff into here.
04:25Anything that you could that is readable by an agent on your computer. So Claude Code, Claude Cowork, and I'm gonna show you how to set this up and get working with it in Claude Cowork as we go through this video as well. Anything that is readable by an agent like Claude Cowork or Claude Code, you'll be able to put into this raw folder for it to read and synthesize into your second brain knowledge base.
04:48Then we have this second step where the AI automatically builds your brain and either do that on a schedule or do that when you trigger it to do it, which is very very simple. You're basically telling the AI to document all of your raw files that are going into the second brain instead of you documenting it yourself.
05:06So you don't have to maintain any of this system really, . That is one of the key parts of this simple second brain setup. Built on the foundation from Andrej Karpathy here, where we're basically dumping in a load of files and we're automating the creation of this second brain, this knowledge base that we can then use to query and work on our own projects.
05:26And then we do one of two things here. We can basically extend the capabilities of our second brain by adding things like automations on top of it, by adding new folders, by adding new prompts, and by adding new tools inside of that second brain to do work for us as well. And you can also use it as your thought partner for either projects or your business.
05:47And those are kind of the two things I would use to set up this simple brain for basically to any projects where you need to have its own brain to kind of use it in that capacity or to set up a second brain for your business so you're automatically documenting all of your business processes and documents in this single place as well.
06:06And if you're looking for a startup idea to build, I do think there's gonna be a cloud version of this sort of system that's gonna exist and be incredibly successful as a startup. So go and build that if you wanna build a cloud based app that is a second brain that companies can kind of plug into and access from anywhere.
06:22I think that idea is gonna be massive in itself. But for now, I wanna show you how to do this locally so that you can set up your second brain and start to use it to improve your work and to access information and do research based on your own life, your own business, your own projects super super easily. And so how do you actually set up this second brain?
06:42Basically, you just need this folder structure here and we have a few folders to call out here. It's five folders and three files. That is literally it to start off with.
06:52You have one folder called raw where you dump all of your files. You have one folder called wiki, which your agent curates all of the raw files and puts them into a knowledge base. You have one folder called archive, which is where all of your old raw files go into once they've been curated.
07:08You have one folder called prompts where all of the prompts for your automations will live. So you can use whatever AI model or automation tool you want to actually run those automations. And you have a projects folder where you can keep all of your individual projects if you're using this in a business capacity to set your second brain up for a business.
07:27And so to break that down just very, very simply, AI is gonna be automatically creating your brain by taking files from raw to Wiki. AI is gonna be archiving our old raw files by taking them from raw to archive.
07:40We're gonna be storing reusable prompts in this prompts folder. I don't even spell that right. You're gonna be documenting your own projects inside of this project folder, and we have a few files here on top that tells the brain basically how to behave and what it can and can't do.
07:55And these are Claude dot m d and agents dot m d. And if you're not familiar with Claude dot m d or agents dot m d, these are basically instruction files that Claude or any other AI agent will run before running the prompts that you give it. So this Claude .md file or agents dot m d file to start off with is basically just gonna have like a rough outline of your goals, business, a bit about you, and your tone of voice.
08:21And you can add these or you can just leave this blank and let the AI automatically create your call dot n d or agents dot n d. We're just creating a very, very simple system here to get our second brain up and running where we're gonna where we have all of these things working together. And all that you have to do as a user really is just dump things into raw and then set up the automations for this.
08:44So now that we've gone down how to actually set up your second brain here, let's jump into Claude Cowork and I'll show you how to set it up in Claude Cowork. A lot of people are working in the Claude desktop app, I wanna show you how to do it inside of Claude. Now this video is sponsored by Anthropic, but I use Claude co work every single day for working in projects in this exact same way.
09:04And so I wanted to show you how to do it and how I use this second brain setup and how to get it up and running for yourself. So you can see here inside of this finder window that I've got. I've got my simple brain and basic setup here, and there is gonna be a link to a GitHub repository with all of this folder and file setup so that you can duplicate it, download it, and create your own second brain from scratch.
09:25There's no complicated complicated install install instructions or no crazy amount of steps that you have to go through. All you have to do is download the simple brain of files and folders from GitHub, and then unzip that, and you'll be able to access that straight away. And I'm gonna show you how to how to start using it with Claude Cowork as well, and how to see some of the files inside of this second brain.
09:45So this is Claude in the desktop app, and all you have to do to use Claude Cowork. If you haven't used Claude Cowork before, is go from chat into this Cowork tab in the middle here on the left, in the left hand navigation. What you're gonna then do is go to projects.
09:58Now I work with this pretty loosely sometimes, so I'm not even using projects for a lot of stuff in Claude CoWork. But if you wanna set up a project here, which is how I recommend getting started with this simple brain, second brain setup is to do a new project here in Claude Cowork. Click on use an existing folder.
10:15Go to select a folder. Select the Simple Brain folder here. We can leave the instructions space blank here as well and then just click create.
10:25And that is gonna create this project inside of SimpleBrain, basically inside of Claude Cowork. And you can see here on the right hand side, we've got context on your computer and we've got our folder structure on the right hand side here.
10:38We can go in and see any of these Markdown files inside of the Claude desktop app here. So this is the read me and we just to go through these files that we've got here. So we've got our simple brain folder with our archive, our projects, prompts, raw, and more wiki folders inside, and then we've got a just m d, cloud dot m d, and read me dot m d.
10:58Inside of the read me, we basically have just some instructions for how this second brain works. So I dump raw stuff into raw, AI turns it into clean notes in Wiki, and that's literally it to start off with. And you can evolve this read me as you go.
11:12So this is just a very basic starting point for creating a second brain. And the reason I wanted to do this is because I think so many people are talking about their crazy complicated second brain setup that they've got. And I just wanna give people I just wanna give you a super super easy way to get started with building a second brain, not to give you the end version of it, which is this crazy complex thing with all of these notes and interlinked stuff in.
11:36And so this is just making keeping it super, super simple, making it really, really easy for you to get started, and this is something that I use for a lot of projects as well. And it's just a really easy way of creating a you create a new brain for a project, you create a new brain for a business, we build all your context into that, and then you use it to kind of query and extend on top of that, and even build a bunch of functionality that you might be using software for into that second brain as well.
11:59So one thing that I use my second brain for is basically taking my latest YouTube videos. So automating, looking at my latest v YouTube videos, taking those YouTube video links, turning those into transcripts, storing that in the raw folder.
12:13That then gets translated into a wiki, and those those transcripts then get put into the archive. I can then start to build or create other content and even layer automations on top of that inside of that second brain project. And I do that across a bunch of different things inside of my business for build great products as well.
12:31So this also outlines the folders that you've got here and then loop. Basically, or drop anything into raw into the raw folder and then run an AI agent with the prompt and translate dot m d. I'm gonna show you that translate prompt as well.
12:43And then read the updated wiki so you can go in and read all of the files that the AI has created. The whole folder is a git repository, so nothing is ever lost. So the idea of this is you don't have to do this when you start off with, but this works with git.
12:57So if you initialize a git repository in this project and basically, what git is is git is version history for local files and folders. This is pre mostly used for code projects, but you can absolutely use it for any type of local projects on your on your computer as well. That is gonna what gonna be what stores the version history of your second brain.
13:18You can do that with GitHub as well and to kind of keep your version history on GitHub if you wanna keep that in the cloud rather than keeping that locally. And that's a really, really good way of if you wanna have your businesses, your company's second brain in a GitHub repository in the cloud that all of your team can kind of access and make updates to as well.
13:37So that's the read me file here for this project. We've also got agents and claw tool m d. And basically, these files and how you can use these files is these are instruction files for the agent, for either Claude or for any other agent.
13:51So codecs or any agents inside a cursor or any kind of agent tool that you're using here, basically instructs it with how it should behave and what it can and can't do. So what I've done here is basically to leave this to be fully customizable so that you can actually customize this for yourself when you download it from GitHub.
14:09It has about the user, and so you're gonna customize this stuff with some basic things about you, some common tasks that it can do. So it's gonna translate the raw stuff into this into the wiki in the wiki folder. You can do a project digest as well, basically summarize a project's name into its read me, and then answer questions.
14:27Basically, read the wiki and archive and answer ad hoc questions about my own past thinking. And there's some rules in here as well, so it's not gonna delete anything from raw or archive. It's not gonna overwrite anything in the wiki blindly.
14:40It's always gonna read it first and then merge things. It's never gonna modify your archive files because that is just old archive files that are gonna be moved into that folder, and it's gonna commit to git after any meaningful changes. Basically, that's just gonna be a save point.
14:54That translates to being like, it's gonna save its process its progress as it goes. And then when uncertain, it's gonna log it in the entry rather than guessing as well so that you know where the AI is kind of making some decisions that you might not entirely want. You've also got a section here for the Wiki voice and structure, so you can customize your tone of voice for your Wiki.
15:15And it's also got a bunch of stuff which is out of scope. But again, these files will be edited as you start using your second brain. This is gonna evolve with you.
15:23This this is not a set file that has to stay exactly this way. And agents dot m d is exactly the same, just some instructions.
15:32The only other file inside of this folder structure here is inside of prompts, and that is called translate dot m d. And this is basically a reusable prompt for your AI agent to read everything in raw, update or create matching markdown files in Wiki so that the Wiki reflects what's in raw, then move each processed file from raw to archive.
15:52The rules is one topic for a file in Wiki. Merge into existing entries, link related entries, move processed files from raw to archive, and never delete them.
16:01If a raw note relates to a folder in projects, also surface it in that project's notes. And when uncertain, note it inside the entry rather than guessing. So this is basically just the instructions for your AI agent to turn your raw files into that wiki that you've got that documents your entire thought process.
16:20And that is the basic setup for the simple brain second brain. And so let's actually start creating stuff inside of our simple brain second brain now and show you a little bit about how it works. So what I'm gonna do to start off with here is I'm just gonna share a few pieces of information to Cowork and drop to get it to drop it into the raw folder, and then we're gonna ask it to run the translate prompt in order to start building our wiki inside of our second brain.
16:48So when you drop stuff into the raw folder here, you can either drop files and You can drop CSV files. You can drop JSON files, document files, PDFs, or other markdown files, or even HTML files from a website.
17:03Or if you wanna source your raw information from somewhere else, can just ask Claude Cowork here to add it to the raw folder for you. So that's what we're gonna do here.
17:12So the first thing I'm gonna do here is I'm gonna add in my business website for build great products. So let's say, can you read this website and capture all of the information about this business in a markdown file in the raw folder? So we're gonna send that to Claude Colerc here.
17:27Let me close this panel here so that we can see Claude working in this session. And what this is gonna do is basically it's gonna it's gonna read build this build great products website, which is my website for build great products. It's gonna find out all of the information on that website.
17:41It's gonna put that into a markdown file inside of that raw folder, and then we're gonna run the translation prompt here to turn that into our wiki inside of this project in Claude Cowork, which we have visibility of all of these files in. If you go back to this project view and click on SimpleBrain, you have all of the visibility of these folders and files here, and it's gonna drop it into this raw folder.
18:04So you can see it's already creating this buildgreatproductswebsite.md, and we've got our outputs listed here in this project as well. And if I jump into this chat, we can see that it is basically outlining all of the information about this business.
18:22So I'm the host. The tagline is helping ideas people build great products with AI. The next generation of incredible companies will start to buy ideas people who build with AI.
18:30We've got all of this information about my communities and my YouTube channel, all of my social links, and a bunch of other notes on the visual and the brand identity as well. And so what we're gonna do now that we've got this markdown file in our raw folder is I'm gonna go back to Simple Brain here.
18:45I'm gonna start a new chat. I'm just gonna say, run the translate prompt in the prompts folder. And then what Claude Cowork is gonna do here is because I'm gonna take that information from the raw folder about the build great products business.
18:59It's gonna turn that into a documented knowledge base inside of the wiki folder, and it's gonna move that markdown file into our archive so that we can start by dropping new files into the raw folder again tomorrow or next week or whenever we wanna do it next. So we can see here it's got instructions. It's got this translate dot m d file that it's read here.
19:20It's got the read me which is read through as well, and it's got the build great products website markdown file and the instructions from Claude for m d that's using all of these things together to start to create this Wiki for us. So you can now see we've also got this build great products dot markdown file, and we've also got this telescope limited dot markdown file and the AI product finder dot markdown file as well.
19:43And these are all and the AI app academy markdown file, they're all being created now. And these are all files that are being created from this markdown file about Bill Gray products that we had in the raw file.
19:54And if we go into the folder here in SimpleBrain in the Wiki, you can see we've got all of these different markdown files for this knowledge base about the different things associated with this business. So you have the AI Academy, which is a community arm of build great products hosted on school, build designership real apps with AI and launch to real customers.
20:12We also have buildgreatproducts.md, which is a outline of the build great products brand. So brand and platform run by Chris Ashby.
20:20I'm helping ideas, people build great products with AI. Here are the offerings, and here is the brand. And you can see how, in a just a few minutes, we're getting set up with a complete knowledge base of our business just from this raw information that we're dumping in the raw folder.
20:35And now what you can do on top of this, because you don't wanna have to always run this translate prompt every single time, is we can start to set up automations that layer on top of this as well to make creating this second brain super, super easy. All we can say here is if we go to this side by a and go to scheduled in Claude, we can then do a new scheduled task, Click set up manually, and then we're gonna go we're gonna call this daily translate.
21:03We're gonna click work in a project, and we're gonna go to work in this simple brain project here. Always allow permissions for that. We can choose the model here as well.
21:10So I'm gonna go with Sonnet 4.6 because we don't need two intelligent model model for this. I think Sonnet 4.6 will do a great job, that will help cut down costs for this as well. The description for this is going to be run the translate prompt in the prompts folder.
21:27Use the translate prompts in the prompts folder to translate raw files from raw into the wiki.
21:37And then we're gonna set this to daily. And you can do it weekly if you wanted to as well. You can just run this every single day, and we click save.
21:44And that is gonna run every day at 9AM inside of the simple brain project. So we now have automated our our translation of raw files into our wiki and this is going to evolve over time.
21:58This is going to grow and build up a complete knowledge base of your entire business or project. And you can see in this archive folder here, you've got the oldbuildgreatproductswebsite.md as well.
22:09And so now I've showed you how to set this up with Claude CoWork, and you can do the same with almost any other agent tool as well. I wanna jump back into the whiteboard and show you the next steps here, what you can do on top of this second brain to start automating it and extending the functionality of your second brain to make it reach those kind of crazy complex levels that you might have seen other people doing.
22:32So the next step on top of your second brain here is to basically extend it with automations. And as we did with adding in the scheduled task there in Claude Cowork, you can do this for a lot of things on top of on top of your second brain. You can start to extend the functionality in a bunch of different ways in order to fit your own project or your own business.
22:53You don't have to have one second brain set up. You can do one for your business, one for individual projects, and do it that way. And especially if you're using this more simple approach, that's gonna give you an easier starting point to get into setting up this simple brain in a way that it's actually gonna work and be way less difficult to maintain over time as well.
23:09So some ideas for some automations that you can do on top of your second brain setup is this scheduled translation. So obviously, we set this up already where you basically set up an automation to run the translate prompt every single day. That means you stop thinking about translating your raw files into your wiki.
23:27It just does it automatically for you. You could set up a voice memo to transcript automation here as well, where you set basically Claude Cowork to watch a folder or basically to look in a folder every single day, take any audio files from that folder, transcribe them with something like Whisper from OpenAI, and then write them to a voice transcription inside of the raw folder to be processed into your wiki.
23:52And so a use case for that might be that you want to just record voice notes on your phone as you're out walking the dog or having your breakfast or whenever any new ideas or inspiration strikes you, just record a voice note that gets saved into a folder. That folder is then observed by Claude Cowork and then those get automatically processed into your wiki.
24:11You could create an automation around YouTube videos and podcast transcripts, basically to check for the latest videos on a YouTube channel every single day or every single week, and then automatically take those, turn them into transcripts using something like YT and DLP, which is basically a tool that you can install inside that project.
24:28If you just ask Claude Cowork to do it, it will just do it for you, and then turn that into a transcript. And again, turn that transcript into your knowledge base automatically using that daily automation as well. And then weekly synthesis.
24:41So you could create a prompt inside of your prompts folder, which basically says, read all of the wiki changes from the last week, and then basically summarize the themes, decisions, and open questions from all of the additions to the Wiki over the last week and create a summary for you. So you have like a weekly or a daily summary, however often you wanna do that, from all of the information that's going into your knowledge base that is evolving over time.
25:05And that is how to set up your simple second brain. How to get started building a second brain if you're new to this and you're getting confused by all of these complicated second brain setups or you've tried a bunch of second brain setups from other people and they just seemed overkill or like way too complicated. This is a very, very simple way to get started to start building a second brain using AI for you or your business that actually works.
25:31So hopefully that helps you understand how you can set up your own second brain if you've been struggling or if you've seen all of these complicated setups or tried something and it was just way too complex complex and you couldn't figure out how to do it or if you're just getting started with trying to figure out how you use AI for your business as well, this can be a really great starting point to kinda create that knowledge base that all of your team can access and add to especially if you're using GitHub as a source of truth and to manage that kind of version control with it as well.
25:59And if you wanna get started with this simple second brain setup, you can either visit my GitHub, which is linked in the description down below, or you can join my community and get access to all of these files and folders as well, which is over on school.com/aiapps, where I've helped thousands of people learn how to build with AI, launch real products, and get real customers for them as well.
26:19If you enjoyed the video, don't forget to like and subscribe. Thank you for watching, and I'll see you next time.
The Hook
The bait, then the rug-pull.
Every second brain video promises a system that changes how you think. Most deliver a maintenance burden that changes nothing. This one argues the opposite — that the right setup is the one you never have to touch.
Named ideas worth stealing.
06:39model
SimpleBrain Folder Structure
- raw/
- wiki/
- archive/
- prompts/
- projects/
- CLAUDE.md
- agents.md
- README.md
Five folders and three files that form a complete self-maintaining knowledge base. AI reads raw, writes wiki, moves files to archive.
Steal forAny project or business that needs persistent AI context without manual documentation overhead.
07:10list
The Three-Step Loop
- Drop files into raw/
- Agent runs translate.md — wiki updated, raw archived
- Query wiki or build automations on top
The repeating cycle that makes SimpleBrain self-compounding over time.
Steal forAny agentic workflow that needs to convert unstructured input into a structured knowledge base.
16:00concept
translate.md Prompt Pattern
- Read everything in raw/
- Update or create matching markdown files in wiki/ (merge, never overwrite blindly)
- One topic per wiki file
- Move processed files from raw/ to archive/
- If raw note relates to a project folder, surface it in that project notes
- When uncertain, note it in the entry rather than guessing
- Commit to git after meaningful changes
A reusable prompt file that turns any AI agent into an automated knowledge base curator.
Steal forDrop this into any Claude project to get automatic wiki generation from any raw inputs.
How they asked for the click.
VERBAL ASK
26:06link
“you can either visit my GitHub, which is linked in the description down below, or you can join my community and get access to all of these files and folders as well”
Dual CTA — free GitHub download plus paid community. GitHub link removes friction for skeptics; community upsell is soft and secondary.
MENTIONED ON CAMERA
FROM THE DESCRIPTION
PRIMARY CTAWhere the creator wants you to go next.
OTHER LINKSAlso linked in the description.
Visual structure at a glance.
