Featured

Deploy OpenClaw in 60 seconds — 20% off logoDeploy OpenClaw in 60 seconds — 20% off

Launch OpenClaw on Hostinger in about 60 seconds and keep your agent live 24/7. Our referral link gives you 20% off, no coupon code needed.

Launch on Hostinger
Run your Hermes agent on Hostinger, fully managed logoRun your Hermes agent on Hostinger, fully managed

Launch Hermes on Hostinger in one click, fully managed, no VPS knowledge needed. Use code ZACAARON10 for 10% off.

Launch on Hostinger
Crawl and scrape any site into clean data, 10% off logoCrawl and scrape any site into clean data, 10% off

Firecrawl crawls and scrapes any site into clean markdown for your agent. Get 1,000 free credits, and new users get 10% off their first purchase.

Try Firecrawl free
Your own AI agent, running 24/7 with QwikClaw logoYour own AI agent, running 24/7 with QwikClaw

QwikClaw sets up and runs an always-on OpenClaw agent for you. One click, no config files, no server setup.

Deploy now
One API to scrape, enrich, and extract the internet. logoOne API to scrape, enrich, and extract the internet.

Context.dev gives your agents a single API to scrape, enrich, and extract live web data — no proxies, no parsers, no maintenance.

Start building free
SetupClaw: done-for-you OpenClaw for founders & exec teams logoSetupClaw: done-for-you OpenClaw for founders & exec teams

White-glove OpenClaw for founders and exec teams (4–50+ employees): we install, harden, integrate your tools, and maintain it — secured from day one.

Get it set up for you
SEO data APIs for your agent, $1 free credit logoSEO data APIs for your agent, $1 free credit

DataForSEO gives your agent live access to SERP results, keyword data, backlinks, and on-page SEO data through one API. New accounts get a $1 credit, good for up to 20,000 keyword or backlink lookups.

Try DataForSEO free
Reach 47,000+ AI builders

A flat monthly placement in front of developers actively installing AI tools. No lock-in, cancel anytime.

Advertise here

Works with

Claude CodeClaude DesktopCursorVS CodeClineCodex CLIOpenClaw+ any MCP client

Install to Claude Code

This server doesn't publish a one-line install command. Follow the setup in the source repository.

Summary

Enables reading and writing MindNode mind maps directly by parsing their on-disk format, without AppleScript or Shortcuts.

README.md

mindnode-mcp

![License: MIT](LICENSE) ![Python 3.11+](https://www.python.org/)

English | 日本語

An MCP server that lets an AI assistant (Claude, etc.) read and write your MindNode mind maps directly — by parsing the .mindnode file format itself. No AppleScript, no Shortcuts, no export/import dance.

Ask your assistant to "read my project map", "add these three ideas under Marketing", "connect the API node to the Auth node", or "turn this outline into a new mind map" — and it operates on the real files MindNode syncs.

Why this exists: MindNode dropped its AppleScript dictionary, and its Shortcuts/URL-scheme automation is thin. But a .mindnode document is just a package whose contents.xml is an Apple binary plist — a clean recursive node tree. Read/write that, and you get full programmatic control.

Requirements

  • macOS with MindNode installed (format version 9 —

current MindNode releases)

Install

git clone https://github.com/masamitsu-konya/mindnode-mcp.git
cd mindnode-mcp
uv sync

Register it with Claude Code (user scope = available everywhere):

claude mcp add --scope user mindnode -- uv --directory "$PWD" run mindnode-mcp

Then start a new session and run /mcp (or claude mcp list) to confirm it shows mindnode ✔ Connected.

By default it finds your MindNode documents in the iCloud container automatically. Point it elsewhere with the MINDNODE_DOCS_DIR environment variable (a local library, or a fixture folder for testing).

Usage

Talk to your assistant in plain language — it picks the right tool. Examples:

| You say | What happens | |---------|--------------| | "List my mind maps" | list_documents → names + dates, newest first | | "Read my Project Plan map" | read_document → the full node tree as JSON | | "Search all my maps for 'pricing'" | search_nodes → every matching node + its document | | "Add 'Hire a designer' under the Team node in Roadmap" | add_node | | "Connect 'Frontend' to 'API' with the label 'calls'" | add_connection | | "Tag the 'Launch' node as #urgent" | add_tag (creates the tag if new) | | "Mark 'Ship v1' as done" | set_task | | "Attach ~/Desktop/wireframe.png to the Design node" | attach_image | | "Make a new map 'Q3 Goals' with branches Sales, Product, Hiring" | create_map |

Nodes can be referenced by their text (a case-insensitive substring is enough) or by their exact id (ids come back from read_document).

Tools

| Tool | Kind | What it does | |------|------|--------------| | list_documents | read | All .mindnode files, newest first | | read_document | read | Mind maps as {id, text, note?, task?, tags?, attachment?, children?} trees, plus connections and the document's tag list | | search_nodes | read | Substring search over node text + notes, in one document or all | | add_node | write | Add a node under a parent (by id or text), with optional note | | add_connection | write | Cross-link two existing nodes with an optional label and arrow direction | | add_tag / remove_tag | write | Tag / untag a node (tags are document-wide and auto-created) | | set_task | write | Turn a node into a checkbox task and set done / todo | | attach_image | write | Attach a local image to a node (copied into the package's resources/) | | create_map | write | Create a new .mindnode from a title + (optionally nested) outline |

Connections (cross-links)

Free links between any two nodes, independent of the parent/child tree (stored at canvas.crossConnections[]). add_connection(document, start, end, label?, direction?)directionforward (default) / backward / both / none. read_document returns each as {id, start_id, end_id, start_text, end_text, direction, label?}.

Tags, tasks, attachments

read_document surfaces these per node (and lists all tag names at the top):

  • Tags — normalized: canvas.tags[] defines {tagID, name, color},

node.tags[] references tagIDs. add_tag auto-defines a tag of that name if it doesn't exist, then attaches it (idempotent).

  • Tasksnode.task = {state, uuids}, where state 1 = todo, 2 = done.

set_task(document, node, done) toggles it.

  • Attachments (images) — `node.attachment = {fileName, size, tintKind,

type}; image bytes live in resources/<fileName>. attach_image` copies the file in and links it, clamping display width to 300px (matching MindNode).

How it works

A .mindnode document is a package directory. Its contents.xml — despite the extension — is an Apple binary plist holding the mind map (format version 9):

canvas.mindMaps[].mainNode        # root node of each map
  ├─ nodeID                       # UUID
  ├─ title.text                   # the node's text, stored as small HTML
  ├─ note / task / tags / attachment
  └─ subnodes[]                   # children (same shape, recursive)
canvas.crossConnections[]         # free links between nodes
canvas.tags[]                     # tag definitions

The server reads and writes this with Python's standard plistlib, so it needs no third-party plist libraries. Node text round-trips through a minimal HTML encode/decode (with proper escaping).

Write safety

These tools mutate your real files, so every write:

  • backs up contents.xml to a timestamped .bak-* first,
  • writes to a temp file then atomically replaces it (no partial writes),
  • preserves keys it didn't author (styling, layout, print info), and
  • drops the stale QuickLook preview so it regenerates.

create_map clones an existing document as a structural skeleton (keeping all opaque auxiliary files valid) and overwrites only the node tree.

Caveat — document open in MindNode. Writes go straight to disk. If MindNode (or another device via iCloud) has the same document open, its next autosave can clobber the change, or you may get an iCloud conflict copy. Close the document in MindNode before writing, or reopen it afterwards to pick up the edit. Backups make this recoverable, but it's cleaner to avoid.

Development

uv run python tests/smoke.py

The smoke test exercises reads against your real documents (read-only) and all writes against throwaway temp copies — it never modifies your actual maps. It also asserts that generated structures (nodes, connections, tags, tasks, image attachments) match the schema of real MindNode files key-for-key.

Status & roadmap

  • [x] Read — list / read / search
  • [x] Write — add_node / create_map
  • [x] Connections / cross-links — read + add_connection
  • [x] Tags, tasks, image attachments — read + write
  • [ ] Non-image attachments (links, stickers)
  • [ ] Tag color palette / rename, task removal
  • [ ] Connection waypoint editing

License

MIT © 2026 Masamitsu Konya

See related servers & alternatives →

Related MCP servers

Browse all →

Related guides

Hand-picked reading to help you choose and use Maps & Location servers.