Author: AI Edge; Compiler: Peggy, BlockBeats

This article introduces a personal knowledge system based on Claude Code and Obsidian. Its core is no longer the usage of "each query and temporary retrieval" in the traditional RAG mode, but an attempt to let AI continuously build and maintain an evolvable knowledge base (Wiki).

Structurally, the system can be split into three layers:
·The first is the original data layer, including unmodifiable input sources such as notes, articles, and transcripts;
·Second, it is a structured knowledge base maintained by AI, which completes cross-reference and relationship construction in continuous updates;
·The third is the Schema rules layer, which is used to standardize the organization of knowledge and the system operation logic.

Around this structure, the system operates through three types of core operations: Ingest, which continuously incorporates external information into the system; Query, which enables instant access to knowledge; and Lint, which checks structural consistency and repairs potential problems.

Under this mechanism, knowledge no longer remains as a one-time conversation result, but gradually becomes a reusable long-term asset through the cycle of "writing-organizing-reusing". Based on this, the author proposes that this model enables knowledge to have an accumulation effect similar to "compound interest": on the one hand, it reduces the individual's cognitive load, and on the other hand, it improves the accuracy and contextual consistency of the model output.

But the effective operation of this system is also based on a premise - continuous input and maintenance. Without stable data injection and structural updates, this "second brain" will be difficult to form a real cumulative effect, and its advantages will also be weakened.

The following is the original text:

Claude Code + Obsidian is the most powerful AI combination I have ever used.

I have almost built an "AI second brain" that incorporates all my thinking, reading, writing, online research, etc. It contains my business plan, all the YouTube videos I’ve ever posted, articles I’ve written, and everything else that’s important to me.

Claude Code + Obsidian has become an instant hit across platforms, and it’s no accident.

For me personally, this AI system has greatly reduced my cognitive load, allowing me to focus on what really matters - whether it's business or my personal life.

My Claude Code + Obsidian System

This system may look complicated, but it actually only takes 5 minutes to set up. More importantly, it has a built-in memory mechanism that will continuously optimize itself with use.

Next, I will take you step by step to replicate this "AI Second Brain" system, which can indeed improve your efficiency.

I recommend you read to the end of the article - I will include a complete cheat sheet for Claude Code + Obsidian, as well as all the resources mentioned in the article (all free).

Before you begin

This system is not my own creation. It was inspired by a popular tweet about "LLM Knowledge Base" by Andrej Karpathy a few days ago.

Related reading: https://x.com/karpathy/status/2039805659525644595

This tweet went viral because it provided an idea for solving a key pain point in the current development of AI.

The problem is this: every time you start a new conversation or switch to a new AI tool, you have to repeatedly re-enter prompt words and add context, almost starting from scratch.

After combining this set of system prompt words with Obsidian and Claude Code, this problem can be completely solved, and at the same time, the output quality of AI can be significantly improved.

How does this system work?

The entire system consists of four core modules:

1. Your data: including articles, notes, transcripts, inspirational ideas, etc.

2. Organization method: Automatically complete organization in Obsidian through Claude Code

3. Instant call: You can ask questions to this "database" at any time and get answers

4. Evolutionary memory: The system will continue to become smarter with use

Your AI second brain

What is the real power of this system?

As humans, our cognitive bandwidth is limited. We forget and sometimes have difficulty connecting different ideas, and there is ultimately a limit to how much information we can track and process at the same time.

With this system consisting of four modules, you are actually releasing your cognitive load and handing over the work of "connecting, organizing and understanding information" to Obsidian and Claude Code.

Your ideas begin to be systematically connected. One note can be automatically linked to another note, and you can re-extract, combine and recall these contents at any time through Claude.

Under this structure, your knowledge is no longer fragmented, but a network that can be constantly called and reorganized - with almost no upper limit.

How to build your AI brain in 5 minutes

1. Download Obsidian

Official website: https://obsidian.md/

https://obsidian.md/

2. Create your Vault (knowledge base)

After the download is complete, Obsidian will prompt you to create a "Vault".

You can think of it as a folder on your computer where we store everything and let Claude Code access and manage this data.

The name of this Vault can be set at will - for example, I call it "Obsidian Vault".

Obsidian Vault (Knowledge Base)

This Vault is where Obsidian stores all your data and notes. All content will be saved in the form of MD (Markdown) files.

3. Set Claude Code

Next, you need to configure a way to access Claude Code. For me (and probably for most people), the easiest way is to just use the desktop client.

In the main chat interface, click "Select Folder", then find the Obsidian Vault you just created and select it.

Claude Code: Connecting your Vault

4. Set system prompts (System Prompt)

Once you have selected the folder, the next step is to paste Andrej Karpathy’s system prompt into the main chat box.

You can copy this prompt here: https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f

Your input should look like this:

Claude Code initial input

Tips: If you don’t want to, you don’t have to open Obsidian manually. Just give the MD folder (that is, your Vault) and related data to Claude Code, and it can directly read, write and modify these files - and these contents will be automatically synchronized to your Obsidian "second brain".

5. Build your database

After you enter the above system prompt words, Claude Code will start to ask you for some data sources to initialize and gradually fill your "second brain".

Build your database

You can think of Obsidian as a "blank notebook" - you need to actively enter content at the beginning, and then the database will be gradually established. Content that can be imported includes: notes, CSV files, Markdown/text files, etc.

Some practical suggestions:

·Export data from your existing note-taking tools

·If you use Notion, you can export to a CSV file

·Let Claude (or other large model) compile a piece of information about you to initialize your "second brain"

·Import your existing articles, collections, inspirational ideas, etc. at once - this is the best time to establish initial data, and you can add it at any time later

It should be noted that a database with a large amount of data like mine is not created overnight, but is gradually accumulated and inputted over time.

my database

That’s it, your “AI Second Brain” has been built and ready to run. Next, I will share some advanced techniques to help you use it more efficiently.

Advanced skills (Pro Tips)

1. Obsidian Chrome extension

If you want to add data to your system more easily, just install Obsidian's Chrome extension. It allows you to click "Add to Obsidian" directly when browsing the web and save the content into your knowledge base with one click. This will make the process of building your "second brain" very smooth.

I myself often use this function to collect articles, web page data, research materials, etc.

Example: Using the Obsidian Chrome extension

It should be noted that the data added through extension is initially an "isolated data source".

Next you can tell Claude Code: "I just added [x] to Obsidian, please help me integrate this content into my wiki."

Claude Code will automatically associate this new data with existing content and generate links, making it truly integrated into your "second brain". This is why this tool set is so powerful.

2. Create separate folders (Vault)

Andrej Karpathy recommends using two separate folders (Vault):

·One for work/business content

·One for personal life/goal management

My own experience is that this structure is the clearest and most effective.

3. Practicality

I found that one of the most valuable uses of this system is actually very simple: to make your LLM prompt words more accurate.

When the model has access to your complete personal information, business plan, writing background and other contexts, it can generate high-quality prompts (even "super prompts") that are more "customized" and closer to reality.

Of course, this system has far more uses than these, but if you just want to start with the most practical scenario, I would strongly recommend that you start with "Improving Prompt Quality".

4. Orphans (isolated nodes)

In Obsidian, "Orphans" are data points that are not connected to other notes.

This feature is useful because it helps you:

·Find ideas that have not yet been integrated

·Discover "weak areas" in the database

·Determine which content deserves further expansion or deepening

In other words, it is not only an organizing tool, but also a mechanism to help you discover blind spots in your thinking.

Orphans (isolated nodes)

You can click the "three dots" in the upper right corner to find and turn on the Orphans switch to see which content has not been associated.

Potential disadvantages of this system

We have talked about many advantages, usage scenarios and optimization methods before, so what are its shortcomings? Under what circumstances would you not be suitable for using this system?

1. People who are not used to visualization

One of the core advantages of this system is that it can visually present data. If you don't rely on or are not used to this method, its help to you may be limited.

2. A certain maintenance cost is required

If you are not willing to continuously maintain a database, then this system may not be suitable for you. Although the maintenance cost is not high, if data is not continuously input into the "second brain", it will be difficult to exert its value.

3. Storage usage

All content will be stored locally as Markdown files, which will take up a certain amount of device space. This also needs to be considered in advance.