for details. Any italicized text below are the notes I added to the Claude draft. I use the term “note” and “file” interchangeably in my own text, because each “note” in Obsidian is actually an individual markdown (.md) file that lives inside the Obsidian Vault folder on my mac.
For those of you brave souls ready to follow me down the Obsidian rabbit hole, here's a detailed breakdown of the plugins, scripts, and workflows I've implemented. Fair warning: this gets pretty technical, so buckle up!
Plugins and supporting apps
Omnivore. This “read it later” service integrates with Obsidian via plugin, and has browser extensions for Firefox and Chrome plus an iOS app that lets me send items to Omnivore from my phone/iPad. This has been essential for replacing the Evernote Web Clipper, which is what’s kept me on Evernote for so many years. It works quite well.
Tag Wrangler: Essential for managing and cleaning up tags. I use this to merge similar tags and maintain a consistent tagging system.
Smart Connections: This plugin creates a local server to generate connections between your notes using AI; a related plugin theoretically creates a graph/mindmap view, but I haven’t been able to make that work yet. To set it up:
Install the plugin through Obsidian's community plugins
Navigate to /Applications/ObsidianPlugins/file-organizer-2000-master/app
Run 'npm start' to start the local server Note: I'm still working on automating this startup process. For now, remember to start it manually each time you open Obsidian.
Templater: Used for inserting dynamic content like creation dates into notes. I've set up templates for consistent YAML frontmatter across all my notes.
File Organizer 2000: While I had high hopes for this one (an AI-powered note organizer and tagger) , it's been a bit problematic. It's supposed to automatically sort and tag files, but I've encountered issues with it not functioning correctly. Use with caution and always back up your files first! I found this was less destructive, if also less useful, once I turned off auto-tagging. Now I’m just using it to file my notes in the right-ish folder, though it’s less than perfect at that too. I have a manual Inbox where my new Obsidian notes get created, and then I drag notes from there into the File Organizer inbox folder when I am ready to trigger its note-filing magic. I am still working from a manual Inbox so that I will have a spot for the files I am currently working on. The file organizer feature is mainly helpful for dealing with all my Omnivore clippings.
Xnapper: This Mac screenshot app is my effort to replace the very broken Skitch app that integrates with Evernote, when it works. Xnapper saves screenshots to my Screenshots folder within my Obsidian Vault (once I told it which folder to save to), just like Skitch saves to Evernote. Then the app “Keep it Shot” uses AI to rename any screenshots that appear in that folder. (Theoretically. Their dev team had to bail me out when it kept breaking.) I didn’t like the number of keystrokes it took to snap with Xnapper so now I have a gesture enabled by BetterTouchTool: Dragging across my mouse pad from left edge to middle will give me the Xnapper crosshairs. Xnapper, Keep It Shot and BetterTouchTool are all part of my Setapp subscription.
TBD: Something to add OCR so that Obsidian will search the text in screenshots and other images. I am also trying a few search plugins but I’m not loving this aspect of Obsidian yet.
Workflow tweaks I have implemented so far
A “created by” system for tagging my notes by source: #cbAlex if I typed it in myself, #cbOminvore if clipped to Obsidian by Omnivore, #cbScreenshot if a screenshot. After conferring with GPT about whether to keep all my clippings in their own folder or in topic folders I realized the whole point of this PKM overhaul is to be able to see related notes together, so I decided I wanted my clips in their related topical folders (though for now, I’m keeping screenshots separate). Using these #cb tags is how I can tell which type of file I’m looking at, and because Obsidian is so customizable, I imagine I can ultimately make notes look different depending on source, and organize them in different ways.
An Alfred workflow for importing GPT exports into my ChatGPT history folder in Obsidian. Cleanup and automation still needed.
Python Scripts
I’m also using Python to help with the cleanup process. See
A lot of my manual or python-enabled cleanup steps will need to be automated so they happen on a recurring basis.
Tag Cleanup: Run the tag cleaning script weekly to maintain a clean tagging system.
File Organization:
Move new files into FileOrganizer for processing
Run without adding tags (due to current issues with tag addition)
Manually review and adjust as needed
Backup:
Set up a regular backup system for your Obsidian vault
Remember: Obsidian sync is not a backup solution!
File Deduplication:
Use the file comparison script to identify potential duplicates
Manually review results to ensure no unique content is lost
Metadata Consistency:
Use Templater to ensure all new notes have consistent YAML frontmatter
Periodically run the YAML formatting script on older notes
Technical Skills Required
To fully implement this setup, you'll need:
Intermediate to advanced Python programming skills [so says Claude! but I am just a monkey-see, monkey-do scripter, pasting whatever ChatGPT writes for me into Visual Studio code. Yes, I’ve tried Github’s Copilot but for now I’ve gone back to GPT.]
Basic understanding of YAML and Markdown formatting [still in progress!]
Familiarity with command-line interfaces and npm [all thanks to GPT! again, I know very little]
Knowledge of file system operations and text processing
MOST IMPORTANT: General nerdiness, a willingness to tinker, and the energy to wade through many blog posts, Discord and reddit discussions, and Obsidian forum posts for guidance
Potential Costs
While most of this setup is free, you might incur costs for:
OpenAI API credits if using AI-powered plugins or scripts
Obsidian Sync if you choose to use their official sync service (remember, this isn't a backup!)
Some Obsidian plugins have fee-for-service versions, largely if based on AI — like Smart Connections and File Organizer
Final Thoughts
This setup is continually evolving. I'm still working on:
Automating the Smart Connections server startup
Improving the File Organizer functionality
Developing more robust file deduplication processes
Planning for my Evernote >> Obsidian migration. There are plenty of plugins and guides to assist in the process but it could get messy.
Figuring out the relationship between my Coda notes, which have displaced at least half of what I used to do in Evernote, and my Obsidian vault. It will probably involve syncing selected pages and tables in Coda into my Obsidian vault, using some kind of local server I run on my Mac. (Obsidian is designed for privacy so doesn’t have turnkey integration with Zapier, IFTTT or Make though there are webhook-based workarounds.)
Remember, the key to making Obsidian work for you is patience and persistence. Don't be afraid to experiment, but always, always back up your data first. Happy organizing!
Want to print your doc? This is not the way.
Try clicking the ⋯ next to your doc name or using a keyboard shortcut (
The Nerdy Details
