MCP server for Obsidian
LocalAbout MCP server for Obsidian
An MCP server that works with an Obsidian vault through the Local REST API community plugin. It suits users who want an AI client to read, search and edit notes, and it is used with clients such as Claude Desktop.
Key features
- 01List files and directories in the vault
- 02Read the contents of a single note
- 03Full-text search across vault documents
- 04Insert content relative to headings or frontmatter
- 05Append content to new or existing files
- 06Delete files or directories from the vault
Use cases
- 01Summarizing meeting notes into a new note
- 02Searching the vault for a topic and explaining the context
- 03Inserting content under a specific heading in a note
MCP support
This project is an MCP server that exposes vault file reading, searching and editing as tools to MCP clients.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEMCP server for Obsidian
MCP server to interact with Obsidian via the Local REST API community plugin.
Components
Tools
The server implements multiple tools to interact with Obsidian:
- list_files_in_vault: Lists all files and directories in the root directory of your Obsidian vault
- list_files_in_dir: Lists all files and directories in a specific Obsidian directory
- get_file_contents: Return the content of a single file in your vault.
- search: Search for documents matching a specified text query across all files in the vault
- patch_content: Insert content into an existing note relative to a heading, block reference, or frontmatter field.
- append_content: Append content to a new or existing file in the vault.
- delete_file: Delete a file or directory from your vault.
Example prompts
Its good to first instruct Claude to use Obsidian. Then it will always call the tool.
The use prompts like this:
- Get the contents of the last architecture call note and summarize them
- Search for all files where Azure CosmosDb is mentioned and quickly explain to me the context in which it is mentioned
- Summarize the last meeting notes and put them into a new note 'summary meeting.md'. Add an introduction so that I can send it via email.
Requirements
- Python >= 3.11
mcpPython SDK>=1.1.0,<2.0.0(pinned inpyproject.toml).mcp-obsidiancurrently registers its tool handlers via themcp1.x low-levelServerAPI (@app.list_tools()/@app.call_tool()), which was removed inmcp2.0. Installing with an unconstrainedmcp>=2.0will crash at import withAttributeError: 'Server' object has no attribute 'list_tools'.
Configuration
Obsidian REST API Key
There are two ways to configure the environment with the Obsidian REST API Key.
- Add to server config (preferred)
{
"mcp-obsidian": {
"command": "uvx",
"args": [
"mcp-obsidian"
],
"env": {
"OBSIDIAN_API_KEY": "<your_api_key_here>",
"OBSIDIAN_HOST": "<your_obsidian_host>",
"OBSIDIAN_PORT": "<your_obsidian_port>"
}
}
}
Sometimes Claude has issues detecting the location of uv / uvx. You can use which uvx to find and paste the full path in above config in such cases.
- Create a
.envfile in the working directory with the following required variables:
OBSIDIAN_API_KEY=your_api_key_here
OBSIDIAN_HOST=your_obsidian_host
OBSIDIAN_PORT=your_obsidian_port
Note:
- You can find the API key in the Obsidian plugin config
- Default port is 27124 if not specified
- Default host is 127.0.0.1 if not specified
Quickstart
Install
Obsidian REST API
You need the Obsidian REST API community plugin running: https://github.com/coddingtonbear/obsidian-local-rest-api
Install and enable it in the settings and copy the api key.
Claude Desktop
On MacOS: ~/Library/Application\ Support/Claude/claude_desktop_config.json
On Windows: %APPDATA%/Claude/claude_desktop_config.json
Development/Unpublished Servers Configuration
{
"mcpServers": {
"mcp-obsidian": {
"command": "uv",
"args": [
"--directory",
"<dir_to>/mcp-obsidian",
"run",
"mcp-obsidian"
],
"env": {
"OBSIDIAN_API_KEY": "<your_api_key_here>",
"OBSIDIAN_HOST": "<your_obsidian_host>",
"OBSIDIAN_PORT": "<your_obsidian_port>"
}
}
}
}
Published Servers Configuration
{
"mcpServers": {
"mcp-obsidian": {
"command": "uvx",
"args": [
"mcp-obsidian"
],
"env": {
"OBSIDIAN_API_KEY": "<YOUR_OBSIDIAN_API_KEY>",
"OBSIDIAN_HOST": "<your_obsidian_host>",
"OBSIDIAN_PORT": "<your_obsidian_port>"
}
}
}
}
Docker
You can run the server in a container instead of installing uv/Python locally.
- Copy
.env.exampleto.envand fill in your Obsidian REST API key:
cp .env.example .env
By default OBSIDIAN_HOST is set to host.docker.internal, which resolves to
your host machine from inside the container (works on Linux, Mac and Windows).
- Build the image:
docker compose build
- Point your MCP client at the container. Since MCP speaks JSON-RPC over
stdio, the client must run it with stdin attached and no pseudo-TTY — use
docker compose run --rm -T, notdocker compose up:
{
"mcpServers": {
"mcp-obsidian": {
"command": "docker",
"args": [
"compose",
"-f",
"<dir_to>/mcp-obsidian/docker-compose.yml",
"run",
"--rm",
"-T",
"mcp-obsidian"
]
}
}
}
You can also run it manually to sanity-check the container starts:
docker compose run --rm -T mcp-obsidian
(it will sit waiting for a JSON-RPC message on stdin; Ctrl+C to exit)
Development
Building
To prepare the package for distribution:
- Sync dependencies and update lockfile:
uv sync
Debugging
Since MCP servers run over stdio, debugging can be challenging. For the best debugging experience, we strongly recommend using the MCP Inspector.
You can launch the MCP Inspector via npm with this command:
npx @modelcontextprotocol/inspector uv --directory /path/to/mcp-obsidian run mcp-obsidian
Upon launching, the Inspector will display a URL that you can access in your browser to begin debugging.
You can also watch the server logs with this command:
tail -n 20 -f ~/Library/Logs/Claude/mcp-server-mcp-obsidian.log
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.