My Obsidian vault became a graveyard of old notes

You opened your vault this morning. The graph view took long enough to load that you switched tabs. Search for a word you know is in there returned a wall of results, none of them the right one. You scrolled. You closed the tab. You opened a new note titled 2026-05-24 and wrote nothing in it.

This is what happens when a tool designed for thinking starts demanding more thinking to keep running, and it is not a moral failure on your part.

Obsidian is a strong note app for local, text-based writing. The local-file model, the plugin ecosystem, the keyboard-only workflows: nothing else gives you that level of control over your own writing. What turned the vault heavy was the content you put into it, not the app itself.

What is the Obsidian PKM tax, and how much is it costing you?

The PKM tax is the time you spend on the vault instead of in the vault. Tagging. Re-linking. Maintaining MOCs (maps of content). Fixing broken plugin updates. Pruning. Migrating between themes. Setting up a new template for the third time this year.

The tax does not show up as a single bill. It shows up as a quiet shift: at some point each week you spend more time grooming the vault than writing in it. That shift is usually invisible until you measure it. Power users who track their own time sometimes describe weeks where maintenance hours crept past writing hours without anyone noticing until the log was tallied.

The most expensive version of the tax is plugin churn. The Obsidian community plugin directory is large and growing. Power users typically run many community plugins at once. After each Obsidian release, forum threads catalog the plugins users say they had to update or replace to keep things working. When a plugin author steps away, the people who depended on that feature describe being suddenly unmaintained. You log in to fix one broken sync, end up reorganizing your daily note template, and the writing you came to do never happens.

If the vault was supposed to give you back hours but is taking hours, the system is upside down.

Why does Obsidian search start failing at scale?

Obsidian's built-in search is fast but literal. It looks for the words you type, in the files you have, with regex support and a few operators. That works beautifully at a few hundred notes. For some workflows it works less well once the vault grows large.

The usual response is to add more structure: tags, properties, smart connections, semantic search plugins. Each addition helps for a week, then becomes another thing to maintain. The third-party Smart Connections plugin uses OpenAI embeddings on your notes and adds a per-month API cost on top of your time. The Omnisearch plugin works well for many users, though some report startup delay during indexing on larger vaults. None of these solve the underlying problem: a vault that grew into a save bucket cannot be searched like a writing archive, because the contents are mixed.

Why is the answer almost never another plugin?

Threads on the Obsidian forum describe a pattern that recurs across users: a plugin solves one symptom, surfaces a new one, and adds time to the weekly maintenance window. Each new layer is one more thing they have to babysit when something breaks.

The deeper fix is usually subtraction. Move the captured-but-not-written content out of the vault. Keep the vault for the things you actually wrote. The graph view becomes useful again. Search returns scannable results. The daily note stops being a checklist of grooming tasks.

This is the move that long-time Obsidian users eventually make, even if quietly. The community gradually concluded that a vault is not a library. A library has stacks. A vault should have shelves.

How does dEssence help when the vault becomes a graveyard?

Obsidian and dEssence solve different jobs. Obsidian is the tool you write in. dEssence is memory you don't have to maintain, and it holds everything else.

dEssence is where the captured-but-not-written content goes: links you want to re-read, articles a friend sent, screenshots, Telegram forwards, browser tabs you can't close, recipe pages, half-watched YouTube clips. Save with the Chrome extension, the Telegram bot, or the web app at dessence.ai, whichever surface is closest at the moment. No folders, no tags, no organizing. Save it, forget it, ask for it later.

When you need a captured thing back, you ask in your own words. "That post about vault maintenance someone shared in March." "The Berlin apartment listing from last month." "The thread on Reddit about Smart Connections." The way you'd describe it to a friend who remembers everything. The answer comes back with the source, the original context, and what you saved with it.

Obsidian stays where it belongs: the vault for things you wrote. dEssence holds the rest.

What can't dEssence do yet?

dEssence is in beta. The Pro tier is in flight and pricing isn't locked. There's no native iOS or Android app today; capture happens through the Chrome extension, the Telegram bot, or the web app at dessence.ai. The free tier has a 500-item ceiling. Team or shared lists aren't built yet. If any of those are blockers for you, the gap is real.

Should you ever delete a note in your Obsidian vault?

The Obsidian community has spent years arguing this question. The honest answer most experienced users land on: yes, but selectively. Delete the imports, the failed templates, the daily notes that are pure logistics. Keep the prose, the synthesis notes, the writing you'd be sad to lose.

Most vault-pruning tutorials suggest the same heuristic: if a note has no inbound or outbound links, no body text beyond a heading, and you don't recognize the title, it's almost certainly safe to archive. Run that filter on a mature vault and many notes will often drop out as orphans.

The lighter the vault gets, the more the original purpose returns.

Frequently Asked Questions

How many notes is too many for Obsidian?

There is no official ceiling, but power users start reporting search and graph performance friction once the vault grows large enough, especially with several heavy community plugins enabled. The exact number depends on hardware, plugin load, average note size, and whether the vault is on local SSD or synced cloud storage. The functional answer is: too many starts when you stop opening the vault because finding things is harder than retyping them.

Why is my Obsidian vault so slow?

The most common cause is plugin load combined with vault size. Heavy plugins (Dataview, Tasks, Templater, Smart Connections, Omnisearch) all walk the file index on open and on edit. A vault with many notes and a large plugin set takes meaningfully longer to start than the same vault with a handful of plugins. Disabling community plugins one at a time is the usual diagnostic step.

Should I split my Obsidian vault into multiple vaults?

For most people, no. Multiple vaults add their own friction: separate plugin configs, separate sync endpoints, no cross-vault search. The better move is usually to keep one writing-focused vault and move the captured-but-not-written content (clips, articles, screenshots) into a separate tool built for recall rather than authorship.

Is Obsidian still worth it after the vault gets huge?

Yes, for the writing. Obsidian's local-file model, keyboard workflow, and plugin ecosystem remain very strong for thinking-in-text. The catch is keeping the vault to that job. When the vault turns into a save bucket for everything you read online, the tool that was helping you write becomes the tool you avoid opening.

How should I clean up an overgrown Obsidian vault?

Start with orphan notes (no inbound or outbound links, minimal body text). Archive rather than delete on the first pass, so you can recover anything important. Then look at imports: web clips, PDF annotations, daily notes that are pure logistics. Those are usually the bulk of the weight. Keep what you wrote. Many people find their actual prose is a minority of the vault, with the rest being capture that belongs elsewhere.