Build an AI Personal Knowledge Base with Obsidian and Claude

Most people have knowledge scattered across notebooks, Google Docs, saved articles, and browser bookmarks they will never revisit. A personal knowledge base — sometimes called a “second brain” — consolidates everything into a searchable, connectable system. Adding AI turns it from a storage system into a thinking partner. Here is how to build one using Obsidian and Claude.

Why Obsidian

Obsidian stores notes as plain Markdown files on your local computer. There is no proprietary format, no subscription required for basic use, and no risk of a company shutting down and taking your notes with it. Its graph view visualises connections between notes, and its plugin ecosystem allows deep customisation.

Step 1 — Install Obsidian and Set Up a Vault

1. Download Obsidian from obsidian.md (free)
2. Open Obsidian → Create New Vault → Choose a folder on your computer (e.g., ~/Documents/MyKnowledgeBase)
3. This folder is your vault. Every .md file you create here is a note.

Step 2 — Set Up the Folder Structure

A structure that works well for most people:

MyKnowledgeBase/
├── 00-Inbox/          Daily captures — process weekly
├── 01-Notes/          Permanent notes on concepts
├── 02-Projects/       Active project notes
├── 03-Resources/      Articles, books, reference material
├── 04-Archive/        Completed or inactive material
└── Templates/         Note templates

Step 3 — Install the Smart Connections Plugin (Claude/AI Integration)

1. In Obsidian: Settings → Community Plugins → Browse
2. Search “Smart Connections” → Install → Enable
3. In Smart Connections settings: Add your OpenAI API key OR configure it to use Ollama (local, free)

Smart Connections indexes all your notes and allows you to ask questions about your knowledge base in natural language. This is where AI changes the game.

Step 4 — The Daily Note Template With AI Hooks

Create a template file at Templates/Daily.md:

# {{date}}

## What I learned today
-

## Connections to existing notes
- [[  ]] — link to related notes

## Questions this raises
-

## Tasks
- [ ]

## AI Summary Prompt
<!-- Paste today's notes here and ask Claude:
"Summarise the key insights from these notes.
Identify connections to prior knowledge.
List three questions worth exploring further." -->

Step 5 — Process Notes With Claude (The Weekly Ritual)

Once per week, do a note processing session:

1. Open your 00-Inbox folder
2. For each note, copy the content into Claude with this prompt:

Here is a rough note I captured:
[paste note content]

Please:
1. Identify the 2-3 most important concepts in this note
2. Suggest 3-5 keywords I should tag this note with
3. Suggest existing knowledge areas this note connects to
4. Write a one-sentence "atomic note" that captures the core insight
5. Draft 2-3 questions this note raises for future exploration

Process the output back into your note. Link related concepts. Move the note from 00-Inbox to 01-Notes.

Step 6 — The Knowledge Graph Query Workflow

With Smart Connections installed, you can query your entire vault:

• Open Smart Connections panel (ribbon icon)
• Type: “What do I know about [topic]?”
• The plugin searches your notes and returns the most relevant passages with links

For a more powerful version, use the Smart Connections chat feature and ask multi-hop questions like: “What connections exist between my notes on machine learning and my notes on business strategy?”

Advanced: Claude API Integration Script

For power users, here is a Python script that processes all unprocessed notes in your Inbox folder automatically:

import os, anthropic
from pathlib import Path

client = anthropic.Anthropic(api_key=os.environ["ANTHROPIC_API_KEY"])
inbox = Path("~/Documents/MyKnowledgeBase/00-Inbox").expanduser()

def process_note(note_path):
    content = note_path.read_text()
    if "## AI Processed" in content:  # Skip already-processed notes
        return
    response = client.messages.create(
        model="claude-3-5-sonnet-20241022",
        max_tokens=1024,
        messages=[{
            "role": "user",
            "content": f'''Process this Obsidian note for my knowledge base:

{content}

Return:
1. Core insight (1 sentence)
2. Tags (5-7 keywords, each on its own line prefixed with #)
3. Related concepts (2-3 existing knowledge areas to link)
4. Two follow-up questions for future notes'''
        }]
    )
    processed = response.content[0].text
    # Append AI processing results to the note
    with open(note_path, 'a') as f:
        f.write(f"

## AI Processed
{processed}
")
    print(f"✓ Processed: {note_path.name}")

for note_file in inbox.glob("*.md"):
    process_note(note_file)

The System That Pays Off Over Time

A personal knowledge base is an investment — the value compounds with use. The first month, you have 30 notes. After a year, you have 300+ interconnected notes and the AI can find connections between things you wrote 11 months apart that you had forgotten were related. This is knowledge that would otherwise be lost to your past browser history.

Key Takeaway: Obsidian provides permanent, local storage for your knowledge. Smart Connections and Claude provide the AI layer for querying, processing, and connecting notes. The combination creates a knowledge system that genuinely amplifies your thinking — not just archives it.