Claude Fable + Graphify = Ultimate Second Brain

Building a Second Business Brain [00:00]

• The core concept is building a "second business brain" that consolidates all business knowledge, including offers, marketing strategies (like Meta ads), and operational details.
• This knowledge is stored in a knowledge graph, which allows for intelligent querying and decision support.
• The system aims to eliminate difficult decision-making by providing data-driven insights.

And one of the really cool things about building a second business brain like this, and specifically using a really special GitHub repo that I'm going to be talking about in today's video, and I'm going to show you step-by-step how to build this, what's special about this is that I don't have to make these insanely hard decisions anymore.

Leveraging Graph AI for Efficiency [00:29]

• A key tool for building this second brain is a special GitHub repo called Graph AI.
• This open-source repository is free to download and use.
• Graph AI significantly reduces token costs when referencing large amounts of data within the knowledge graph.

Because what I can do is I can take this second brain of mine using this knowledge graph, and then I can open up something like Cloud Code, for example. And then what I can do is I can start asking questions to my second brain using this system, and I can get some of the best answers that you could possibly imagine without having to sit there and manually really think about anything.

Token Cost Savings with Graph AI [03:06]

• Without Graph AI, processing 321 files through a standard Claude code session would be prohibitively expensive.
• Graph AI offers a more efficient sorting and referencing algorithm, akin to a RAG database setup.
• This efficiency results in an estimated 72% saving on token costs, making queries approximately 3.6 times cheaper.
• For instance, a query costing 62,000 tokens with standard methods could cost only 17,000 tokens with Graph AI.
• Over time, these savings compound, preventing high monthly AI subscription costs, which is crucial for budget-conscious users.

But here is what is so so crazy when we use Graphy in the mix because it has a better sorting algorithm. It has a better setup for referencing information in these huge databases like a rag database or even just a regular code base. It's so much more efficient and it's so much more token friendly or budget token budget friendly I should say.

Setting Up the Second Brain [05:21]

• The first step is to create a folder and a .md file within it.
• This .md file needs a specific structure to properly source the second brain system.
• A prompt for creating this .md file is available for free in the community's YouTube resources section.
• The system automatically generates three main folders: "raw," "wiki," and "output."

And the best way to go about doing that is if I head on over to my free community where you can join completely for free with that first link down in the description. I'm going to post the specific prompt that we used to make that .md file.

Organizing and Connecting Data [06:41]

• The "raw" folder acts as a dump for various file types, including documents, web articles, PDFs, images, and videos.
• The "wiki" folder is where sourced information is organized into specific sections or wikis.
• Graph AI is used to source items from the "raw" folder into their appropriate "wiki" locations, establishing connections and tags.
• Users can instruct Graph AI to audit the knowledge base, ensuring all connections are logical and adding new ones where necessary.

And then what you can do is using graphify, you can then take those images and everything inside of glad code and then source them into their individual places where they live. And in this case, let's say, you know, you have different wikis and those wikis are where they would live and everything has connections and is tagged appropriately using glad code as the agent surface to do that.

Managing the Second Brain [09:04]

• Once set up, the second brain can be managed by interacting with Claude Code as one would any other application.
• The system handles connections and node management automatically through these conversations.
• Users are encouraged to join the free community via a link in the description for resources and templates.

Then, once you have that, you can manage your second brain by just talking to Claude Code like you would any other application there. You can just have conversations with Claude Code, and then it will take care of the rest in terms of connections and looking at nodes and things of that nature.