Vault Cortex — Claude Code MCP server for Obsidian
Connects an Obsidian vault to an AI agent for search, task handling, memory, and file access. Useful when you want an agent to work with notes directly.
Open on GitHubWhat Vault Cortex does
Vault Cortex connects an Obsidian vault to an AI agent over MCP. It searches notes, reads and edits files, handles tasks, and keeps structured memory. Search combines text matches with semantic results; the server also works with Obsidian links, tags, and properties.
You need Docker and an Obsidian vault, or a folder containing .md files. The CLI requires Node.js 22.12 or later. The server reads files from disk, so Obsidian does not need to be open. One deployment detail matters: the CLI and Docker Compose manage containers separately, so use the same method for later updates.
What it helps with
- find a note by meaning when its wording differs from the query
- read or edit a file in an Obsidian vault
- filter tasks by status, date, or priority
- save a dated entry to structured memory
Example prompts for the agent
- Find notes about my previous trip and summarize the route from my vault.
- Show high-priority tasks and move the selected task to another status.
- Save this conclusion to memory with today's date.
How to install Vault Cortex
Install
The MCP registry has no launch command for it: see the repository README
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Vault Cortex
How do I install Vault Cortex for Claude Code?
Install Docker and Node.js 22.12 or later, then run npx vault-cortex@latest init. The setup wizard asks for the vault location and prepares the MCP client connection.
What does Vault Cortex do?
It connects Obsidian to an AI agent through MCP, providing search, file access, task operations, and structured memory.
Does Obsidian need to be running?
No. Vault Cortex works with vault files on disk, so Obsidian can remain closed.
