query_vaultLABSCO SUMMARY
The maintenance tools are the reason to install it; search is table stakes.
Any vault bridge can find a note. What is unusual here is being able to ask which notes arrived recently and never got linked to anything, or which clusters have grown dense enough to deserve a summary — questions you cannot ask a search box. Deliberately uncached, so results are always current and large vaults will feel it.
What it is
A TypeScript server over a local Obsidian vault. Beyond read and write it ships a set of maintenance tools that answer questions about the vault's shape — which notes are unconnected, which clusters need a synthesis note — rather than only about its contents.
What you get
- Reading:
get_notefetches one note,list_directoriesbrowses the structure with note counts,get_backlinksfinds what links to a note with context - Searching:
search_notesdoes exact filename or content matching,intelligent_searchcombines direct matches, link proximity, tag expansion and structural context,query_vaulttakes a plain-language question - Writing:
write_note,create_notewith frontmatter and tags,append_to_note, andupdate_note_sectionwhich replaces the content under one heading - Maintenance:
audit_recent_notesfinds recently modified notes missing required frontmatter fields,fresh_energyfinds new notes with no links either way,synthesis_readydetects interlinked clusters lacking a summary note - Project work:
initiative_bridgetracks notes tagged to a project that still carry open tasks,pattern_echofinds notes reusing a phrase or framework fragment guided_pathbuilds a narrative tour from a seed note, with supporting threads, counterpoints and action items
Requirements
One environment variable, OBSIDIAN_VAULT_PATH, set to the absolute path of your vault. Build from source with npm install and npm run build, then point the client at dist/index.js with an absolute path. Nothing is cached — every search reads the vault live, which avoids stale results at the cost of speed on very large vaults.
Setup effort
Build from source — git clone https://github.com/dbmcco/obsidian-mcp.git
USE CASES
01
Finding notes a keyword search would miss
Search with link proximity and tag expansion so neighbouring notes surface too.
02
Catching notes that never got integrated
List recently updated notes with no backlinks or outgoing links and connect them.
03
Knowing when to write a summary note
Detect clusters of interlinked notes with no synthesis and write the missing one.
04
Chasing open tasks on a project
Pull the notes tagged to that initiative that still have outstanding items.
05
Fixing metadata after a busy week
Audit recent notes for missing frontmatter fields, then update them.
17 tools
Listed in project docs. Your client may show a different list.
search_notesintelligent_searchlist_directoriesget_noteget_backlinkswrite_notecreate_noteappend_to_noteupdate_note_sectionguided_pathaudit_recent_notescontextual_companionsfresh_energyinitiative_bridgepattern_echosynthesis_readyBEFORE YOU INSTALL
What to know first
Tools you get
17 tools named in the project's documentation.
Requires
- Node.js and npm
- An Obsidian vault, whose absolute path goes in OBSIDIAN_VAULT_PATH
Where it runs
Both documented clients run the server on your machine.
INSTALLATION
Connect your client
You clone and build this one yourself.
Server setup
TERMINAL
git clone https://github.com/dbmcco/obsidian-mcp.git
cd obsidian-mcp
npm install
npm run buildMCP Clients
Claude Code, Claude Desktop
JSON
{
"mcpServers": {
"obsidian": {
"command": "node",
"args": [
"<YOUR_PATH_TO_OBSIDIAN_MCP>/dist/index.js"
],
"env": {
"OBSIDIAN_VAULT_PATH": "<YOUR_OBSIDIAN_VAULT_PATH>"
}
}
}
}
