Multiple Obsidian vaults sharing one “brain” via symlinks – plugins not activating in cloned vaults

Help
,
1

Hi,

I’m a macOS user experimenting with a deliberate, non-standard Obsidian architecture, and I’m looking for technical confirmation of what is and is not possible, not tag-based workarounds.

I fully understand tag trees, path filters, and workspaces.
Please consider this question strictly from a filesystem + Obsidian runtime architecture perspective.

Goal

I want to achieve:

  • ONE shared “BRAIN”

    • same community plugins

    • same hotkeys

    • same snippets

    • same templates

    • same UI / workflow

  • MULTIPLE VAULTS

    • each vault = a fully isolated project tree

    • separate:

      • tags

      • graph

      • backlinks

    • clean conceptual separation per project

In short:
same tools, different knowledge spaces

Proposed architecture (macOS symlinks)

Vault-Project-A/
└── .obsidian/  -> symlink to Vault-Brain/.obsidian

Vault-Project-B/
└── .obsidian/  -> symlink to Vault-Brain/.obsidian

Vault-Project-C/
└── .obsidian/  -> symlink to Vault-Brain/.obsidian

Example:

.obsidian/
├── app.json
├── appearance.json
├── backlink.json
├── bookmarks.json
├── canvas.json
├── community-plugins.json (symlink -> /VAULT-BRAIN/community-plugins.json)
├── core-plugins.json (symlink -> /VAULT-BRAIN/core-plugins.json)
├── daily-notes.json
├── global-search.json
├── graph.json
├── hotkeys.json (symlink -> /VAULT-BRAIN/hotkeys.json)
├── page-preview.json                   
├── plugins (symlink -> /VAULT-BRAIN//plugins)              
├── snippets  (symlink -> /VAULT-BRAIN/snippets)
├── templates.json
├── workspace.json
└── workspaces.json

Inside .obsidian, the following are shared via symlinks:

  • community-plugins.json

  • core-plugins.json

  • hotkeys.json

  • plugins/

  • snippets/

  • templates.json

  • other UI config files

At the filesystem level, this works correctly.

Observed result

  • UI is identical across all vaults

  • Hotkeys, snippets, templates are shared

  • Each vault has its own tags and graph

  • NO?! Community plugins are only enabled in the original “brain” vault

In cloned vaults:

  • plugins do not appear as enabled

  • Settings → Community plugins behaves as if plugins were disabled

  • despite community-plugins.json and plugins/ being present and shared

Key observation

It appears that Obsidian does not rely solely on .obsidian/ to determine plugin enablement.

Some additional runtime state seems to be stored outside the vault, most likely in:

~/Library/Application Support/Obsidian/

This likely includes:

  • vault identity

  • plugin enable/disable state

  • binding between plugins and a specific vault ID

This would explain why:

  • config files are shared successfully

  • but plugins do not “clone” across vaults

The actual question

  1. Is it technically possible to force Obsidian to:

    • run multiple vaults

    • while sharing a single plugin “brain”

    • without manually enabling plugins per vault?

  2. Has anyone on macOS:

    • used this symlink-based multi-vault model long-term?

    • encountered hard limitations or corruption risks?

  3. Is this behavior a fundamental architectural decision in Obsidian,
    rather than a limitation of symlinks or macOS?

I’m specifically interested in real-world experience and technical confirmation, not theoretical workarounds.

Thanks in advance for any deep technical insight.

2

There has been discussion about using the same configuration for multiple vaults. Searching the forum should find some. There may be relevant discussion/links in this feature request: Global Settings / Same settings, themes, and plugins across multiple vaults.