# How I use Obsidian + Claude Code to run my life

Source: https://www.youtube.com/watch?v=6MBq1paspVU
Recap page: https://rapidrecap.app/video/6MBq1paspVU
Generated: 2026-02-23T20:32:31.905+00:00

---
## Quick Overview

The integration of Obsidian, utilized as a centralized knowledge vault built on markdown files and inter-file relationships, with Claude Code via the Obsidian CLI enables powerful AI delegation by providing LLMs with deep, contextual understanding of the user's entire thought history, leading to better ideas, automation, and decision support.

**Key Points:**
- Cloud Code acts as an agent in a command line interface that controls the computer using natural language, but its effectiveness hinges on providing it with good context, which is often lost in standard sessions.
- Obsidian functions as a second brain built on markdown files that uniquely tracks inter-relationships between files, and the Obsidian CLI allows Claude Code to read not only the files but also these critical relationship structures.
- Custom commands like /context load full context about my life, work, and current state allow for instant preloading of extensive personal and project information into Claude Code sessions, saving repeated explanations.
- The integration surfaces patterns about the user's thinking that they would not notice alone, exemplified by the /trace demo which showed the 13-month evolution of the user's understanding of Obsidian and backlinking.
- The system supports moving from reflection to action, as demonstrated by the /ideas demo which generated actionable outputs like the suggestion to 'Build the /graduate command' and specific tools to build.
- The host asserts that users serious about LLMs for building a 'personal OS' are not using LLMs properly if they do not use a centralized markdown-based note-taking tool like Obsidian.
- Giving an agent like Claude Code access to the Obsidian vault, which contains deep personal information, is described as 'scary' and raises fundamental questions about the evolution of privacy.

**Context:** The video features a conversation between the host and guest Vin (Internet Vin) about leveraging Obsidian, a tool that manages notes as interconnected markdown files forming a 'vault,' in conjunction with Claude Code, an agent accessible via the command line interface. The core premise is that combining Obsidian's relational knowledge structure with Claude Code's execution capabilities allows users to delegate complex tasks and gain profound insights by feeding the LLM unparalleled, persistent context about the user's life and thoughts.

## Detailed Analysis

Vin explains that Claude Code allows computer control via natural language commands, but users often fail to achieve game-changing results because they do not feed it the necessary context repeatedly. Obsidian solves this by serving as a structured knowledge base where files are connected through explicit relationships, which the Obsidian CLI exposes to Claude Code. This connectivity allows the LLM to understand latent ideas and patterns across the user's entire history, leading to breakthroughs. Vin demonstrated numerous custom commands, such as '/context load full context about my life, work, and current state' to instantly prime the agent, and '/ghost' to answer questions in the user's established voice by building a voice profile from the vault. The most powerful demonstrations involved analyzing the vault's history, like the '/trace demo' which mapped the user's evolving relationship with Obsidian over 13 months, pulling evidence from old notes. The integration moves beyond reflection to action; an '/ideas demo' generated actionable suggestions, including tools to build and specific people to meet, based on synthesizing thousands of notes. While the process requires significant time for setup and execution due to reading extensive files, the host argues that this contextual depth is essential for maximizing LLM utility, even suggesting that inline delegation via notes might become a new UX pattern, though the security implications of granting an AI access to one's entire 'second brain' are acknowledged as a significant concern.

### Cloud Code Capabilities

- Agent control via natural language
- Requires persistent context for complex tasks
- Problematic repetition of context in standard sessions

### Obsidian as the Context Source

- Tool manages markdown files forming a vault
- Uniquely connects relationships between files (not just folders)
- Obsidian CLI exposes file structure AND relationships to Claude Code

### Key Custom Commands Demonstrated

- /context load (loads full life/work context)
- /slash today (morning review using calendar/notes)
- /ghost (answers questions in user's voice after building profile)
- /connect (links two domains using the vault's graph)

### Impact of Contextual Depth

- LLM surfaces unstated patterns and latent ideas
- Trace command reveals historical evolution of concepts (e.g., Obsidian usage)
- Better ideas are generated at the right time, leading to happier, healthier, wealthier lives

### Transitioning from Reflection to Action

- Idea generation report suggests tools to build (e.g., /graduate command) and systems to implement
- Results move from personal analysis to actionable building tasks
- The agent can suggest new commands based on the user's perceived skill level and vault content

### Setup and Usability Concerns

- Custom commands must be created, often by asking Claude Code itself
- Requests accessing the full vault take significantly longer than standard LLM prompts
- The blank canvas UI of Obsidian can be daunting for new users