Turn Obsidian Into
Your JARVIS with
Claude Code

Overview

What's In This Guide

The Insight

Why This Works When Other AI Setups Don't

Every time you open ChatGPT, it starts from zero. You re-explain your business, your preferences, your context. You're constantly prompting a generic AI that doesn't know you.

Obsidian stores everything as plain-text markdown files. Claude Code reads files natively. When you run Claude Code from inside your vault, it can open, read, and reason across every note you've ever written.

No exports. No copy-paste. No re-explaining. Claude reads your actual notes, projects, and ideas. You stop prompting a generic AI and start talking to one that knows your work.

Before You Start

What You Need

Two things: Obsidian (any existing vault works) and Claude Code installed on your computer. That's it. No coding background required.

If you don't have Claude Code yet, install it with one command in your terminal:

Windows users: download the installer from claude.ai/download. Already have Claude Code? Skip ahead.

Step 1

Point Claude Code at Your Vault

Open your terminal. Navigate to the folder where your Obsidian vault lives. Then run Claude Code.

$ cd /path/to/your/obsidian-vault
$ claude

Replace /path/to/your/obsidian-vault with the actual folder path. On Mac, you can find this by right-clicking your vault in Finder and selecting "Get Info."

Claude Code will ask you to approve the folder. Hit enter to trust it. That's the connection made.

Step 2

Write Your CLAUDE.md

This is the file that makes the difference. Without it, Claude is smart but generic. With it, Claude knows who you are, how your vault is organized, and what you're working on right now.

Create a file called CLAUDE.md at the root of your vault. Claude Code auto-loads this file at the start of every session. It's your persistent context file. Write it once and it follows every conversation.

# [Your Name]: Vault Context for Claude

## Who I am
[Your role, what you're working on, relevant background. 2-3 sentences.]

## How this vault is organized
[Folder names and what lives in each one. Example:]
- 00-Inbox/: raw capture, quick notes
- 01-Projects/: active work
- 02-Reference/: research, saved articles
- 03-Journal/: daily notes

## Current focus
[What you're working on right now. Update this every few sessions.]
- [Active project 1]
- [Active project 2]

## Key files Claude should know about
[Paths to notes that give important context]
- /01-Projects/[your main project].md
- /02-Reference/[important reference].md

## What Claude should always do
- [Preference 1, e.g., "always ask before creating new files"]
- [Preference 2, e.g., "use bullet points over paragraphs"]

Step 3

Start Your First Session

With your CLAUDE.md in place, run claude from your vault folder again. Now test it with these three prompts. Each one shows a different layer of what's possible.

Optional Power Move

Run Claude Code Inside Obsidian

If you want Claude Code running side-by-side with your notes without switching to a separate terminal window, install the Terminal community plugin for Obsidian.

Now Claude Code lives inside Obsidian. You write a note, then immediately ask Claude about it. No switching apps.

The Compounding Effect

What This Unlocks

Here's the part most people miss: this gets better over time. Every note you write adds more context. Every session builds on the last. Most AI tools reset. This one compounds.