MCP Note-Taking Server

The MCP Note-Taking Server lets you manage personal notes with structured tagging, markdown formatting, and a web interface. It integrates with Claude Desktop so you can create, find, organize, and export notes during AI conversations with a smooth, wiki-like browsing experience.

How to use

You interact with the MCP Note-Taking Server through your MCP client (for example, Claude Desktop) to create notes, organize them with a structured tag schema, search across titles and content, and export notes to Markdown. Each note stores a title, Markdown content, a set of tags (category, type, priority, topics), and timestamps for created and updated times. Use the tag filters to narrow results, combine a title search with date ranges, and take advantage of the web interface for browsing and linking related notes.

How to install

Prerequisites: you need Python 3.10 or higher and Claude Desktop installed on your machine.

Check your Python version to ensure you have a compatible runtime.

python3 --version

Install the required Python dependencies for the MCP server.

python3.11 -m pip install -r requirements.txt

Configure Claude Desktop to load the MCP server. Add or update your configuration with the MCP server entry that points to your local server script.

{
  "mcpServers": {
    "notes": {
      "command": "python3.11",
      "args": ["/absolute/path/to/MCPNotes/server.py"]
    }
  }
}

Replace the Python command if you use a different Python 3.10+ executable, and replace the server path with the actual location of your server script.

Restart Claude Desktop to load the MCP server. Then verify the integration by asking Claude Desktop what tools are available.

Additional sections

Data model and tag schema

Note structure includes an id, title, content (Markdown), tags, created, and updated timestamps.

Tag schema includes:

• category: exactly one of work, personal, learning (required)
• type: exactly one of project, idea, reference, todo, note (required)
• priority: exactly one of active, soon, someday, eventually, maybe, not-actionable (required)
• topics: zero or more of mcp, ai, coding, design (optional) Notes persist in a JSON file with atomic writes to prevent corruption.

Web interface features let you browse notes in a wiki-like format with backlinks, tag navigation, full-text search, and Markdown rendering. The web UI reads directly from the same storage file used by the MCP server so changes appear in real time.

If you run into issues, verify the MCP command and path, ensure the notes storage file is writable, and check permissions on the directory. Ensure your tag selections meet the required schema to avoid validation errors.

Notes about usage and troubleshooting

Available tools let you manage and search notes from the MCP interface. You can create, read, update, delete, and export notes, as well as view and adjust the tag schema.

If you see tags validation errors, consult the tag schema to confirm the required fields are provided and that values match exactly (case-sensitive). If notes don’t persist, verify that the storage file exists, is writable, and that you have the necessary file permissions.

Available tools

get_tag_schema

Retrieve the complete, current tag schema showing all valid values for category, type, priority, and topics.

create_note

Create a new note with validated tags and Markdown content. Returns the new note id and metadata.

update_note

Update an existing note; supports partial updates to fields including title, content, and tags.

delete_note

Remove a note by its unique id from the storage.

read_note

Return the full details and content of a specified note by id.

find_notes_by_tags

Search notes using tag filters, title containment, and date ranges with AND logic across filters and OR across topics.

list_tags

List all active tags with counts showing how many notes use each tag.

add_tags_to_schema

Dynamically add new values to any tag dimension in the schema (category, type, priority, topics).

export_note_to_markdown

Export a single note to a Markdown file at a specified path.

export_all_notes_to_markdown

Export all notes to Markdown files in a chosen directory.