Obsidian With Claude: The Setup I Said You Didn't Need
The video demonstrates how to set up Obsidian with Claude for AI-powered knowledge management, showing it as an optional but useful interface layer on top of local folders. The creator walks through installing a terminal plugin, configuring Claude to recognize the Obsidian environment, and building practical workflows like journal entries, people databases, and canvases.
Summary
The video begins with the creator clarifying a point from previous videos: Obsidian is not required for AI-powered knowledge management, but it can serve as a useful visualization and access layer on top of local folders. He emphasizes that the goal is to show the simplest possible setup for those who prefer working within Obsidian.
The setup process starts with ensuring Obsidian is on version 1.12.7, which includes the Obsidian CLI — a command-line interface that allows Claude to interact with Obsidian files more efficiently. The creator creates a new vault on the desktop to demonstrate that a vault is simply a local folder, making the point that Obsidian is just an interface over standard file storage.
To bring Claude inside Obsidian, the creator installs a community plugin called 'Terminal,' which embeds a terminal window directly within the Obsidian sidebar. This allows Claude to be launched within the active vault folder automatically, eliminating the need to manually navigate to the correct directory.
Once Claude is running inside the vault, the creator uses the '/init' command to help Claude recognize it is operating in an Obsidian environment. He then instructs Claude to create a CLAUDE.md file that documents the available Obsidian CLI capabilities, ensuring Claude can intelligently use the CLI in future interactions. The creator notes that Claude doesn't always auto-detect the CLI and may need to be explicitly told about it.
With the environment configured, the creator demonstrates practical use cases. He has Claude backfill 30 days of journal entries with fictional people, topics, and backlinks to showcase Obsidian's graph and linking features. He then asks Claude to reorganize these entries into a dedicated journal folder and update Obsidian's settings files accordingly — all without manually touching the settings UI.
The creator shows how to use Claude with the '--dangerously-skip-permissions' flag to bypass per-change approval prompts, and explains the different modes available (auto-accept edits, plan mode, etc.). He also demonstrates session recovery using 'claude resume' after a crash caused by Obsidian reloading.
A key workflow shown is creating journal entries from screenshots. The creator pastes an image into the terminal, and Claude reads it, creates a dated journal entry, saves the image to an attachments folder, and appends it to the daily note. He then improves this by instructing Claude to create a slash command ('journal image entry') stored as a prompt file, enabling one-command journal creation in future sessions.
The creator also demonstrates people management: Claude creates a people database with metadata fields (location, relationship, last contacted), and the creator refines rules so that new people mentioned in journal entries are automatically added to the people folder and database with cross-links.
Finally, the creator shows that Claude can create Obsidian canvas files by writing the underlying JSON format directly, even though canvases are not natively supported by the CLI. He suggests updating the CLAUDE.md file to reinforce this behavior. The video closes with a mention of future content on integrating AI agents into the system.
Key Insights
- The creator argues that Obsidian is not necessary for AI-powered knowledge management — it is simply an additional visualization layer on top of a standard local folder, and Claude can manipulate files effectively without it.
- The creator explains that Claude does not always auto-detect the Obsidian CLI and may incorrectly claim it has no access, so users should explicitly instruct Claude to check for CLI availability if it denies it.
- The creator demonstrates that Claude can modify Obsidian's internal settings files directly through the file system, bypassing the need to use the Obsidian UI for configuration changes like setting the default journal folder.
- The creator shows that slash commands in this setup are not special integrations — they are simply text files containing saved prompts stored in a 'commands' folder, which Claude reads and executes when triggered.
- The creator explains that Obsidian canvases, while not supported by the Obsidian CLI, can still be created by Claude because canvas files are just JSON files that Claude can write directly to disk.
Topics
Transcript
[0:00] If you watch the previous videos, you know that we can build a local folder and use AI for knowledge management, and I made a few videos about why you don't need to use Obsidian to do that. People say, but I love using obsidian with this, and this is great. And this video I will show for those who are interested in using Obsidian exactly step by step, how you set these things up. However, the point was in the previous videos that there is no need. To use Obsidian in order to leverage AI for knowledge management. [0:33] It is another layer on top, and I just had a feeling that people think they need to use…
Full transcript available for MurmurCast members
Sign Up to AccessMore from ICOR with Tom | AI Productivity
Obsidian is the tool I teach in. Keep the one you have.
The creator addresses criticism about switching to Obsidian after previously promoting other tools, clarifying that his tool-agnostic ICOR methodology remains unchanged and that Obsidian is simply an interface layer for his folder-based system. He emphasizes that the underlying folder structure, AI automation, and productivity principles are identical regardless of which tool visualizes the content.
Claude sucks. And here is why.
This video explores why Claude and other AI models produce inconsistent results, examining how model selection (Haiku vs Sonnet vs Opus vs Claude 3.5 Fable), effort levels, prompt clarity, and system context dramatically affect output quality. The speaker demonstrates these differences through a Venn diagram creation task and advocates for using organized folder systems with documented SOPs and code-based solutions to ensure consistent AI performance.
Claude can now prompt its own SESSIONS. I built the boss.
Claude now has the ability to interact with multiple sessions, allowing a streamlined orchestration of tasks through an agent called Larry. This integration can simplify task management and enhance productivity by enabling a single main session to oversee various sub-sessions or tasks.
I switched off Claude's memory. My folder does it better.
Tom explains why he disabled Claude's native memory feature and instead built a local folder-based system (myPKM) to maintain persistent context with AI agents. He shares an audit revealing systemic issues like contradictory documentation, forgotten rules, and parallel session conflicts, then outlines how to systematically fix these problems using the ICOR productivity methodology.
I do not download invoices anymore. My folder and Claude do it.
The speaker demonstrates how he uses AI (Claude) integrated with his local folder system to automate invoice collection for his accountant, downloading 36 invoices in 5 minutes with minimal manual intervention. He explains his AI team methodology, work streams, and how persistent local knowledge bases enable AI to handle recurring tasks independently without constant re-explanation.