A Model Context Protocol server for reading, creating and querying XMind mind maps. This tool provides powerful capabilities for searching, extracting, analyzing and generating XMind files.

• Parse complete mind map structure (multi-sheet)
• Smart fuzzy search across mind maps
• Task management and tracking (to-do + planned tasks)
• Hierarchical content navigation
• Link and reference extraction (external URLs + internal xmind:# links)
• Multi-file analysis
• Label, callout, boundary and summary support
• Directory scanning

• Create XMind files from structured JSON
• Nested topics with notes (plain text + HTML formatting)
• Labels, markers, callouts, boundaries, summaries
• Relationships between topics (by title)
• Internal links between topics across sheets (linkToTopic)
• Simple to-do tasks (checkbox)
• Planned tasks with Gantt support (dates, progress, priority, duration, dependencies)
• Predefined themes (default, business, dark, simple)
• Layout structures (clockwise, logic.right, org-chart, fishbone, timeline, etc.)
• Overwrite protection

• Optionally restricts access to specified directories (if none provided, any path is allowed)
• Path normalization and validation
• Error handling for invalid access attempts

npm install @modelcontextprotocol/sdk adm-zip zod
npm install --save-dev typescript @types/node

node dist/index.js <allowed-directory> [additional-directories...]

1. read_xmind - Parse and extract complete mind map structure
2. list_xmind_directory - Recursively scan for XMind files
3. read_multiple_xmind_files - Process multiple files simultaneously
4. search_xmind_files - Search files by name or content
5. extract_node - Smart fuzzy path matching with ranked results
6. extract_node_by_id - Direct node access by ID
7. search_nodes - Multi-criteria search (title, notes, labels, callouts, tasks)

8. create_xmind - Create XMind files from structured data

{
    "name": "read_xmind",
    "arguments": {
        "path": "/path/to/file.xmind"
    }
}

{
    "name": "search_nodes",
    "arguments": {
        "path": "/path/to/file.xmind",
        "query": "project",
        "searchIn": ["title", "notes"],
        "caseSensitive": false
    }
}

{
    "name": "create_xmind",
    "arguments": {
        "path": "/path/to/output.xmind",
        "sheets": [{
            "title": "Project Plan",
            "theme": "business",
            "rootTopic": {
                "title": "Deployment",
                "children": [
                    {
                        "title": "Analysis",
                        "durationDays": 3,
                        "progress": 0,
                        "priority": 1
                    },
                    {
                        "title": "Development",
                        "durationDays": 5,
                        "progress": 0,
                        "dependencies": [{"targetTitle": "Analysis", "type": "FS"}]
                    }
                ]
            }
        }]
    }
}

{
    "name": "create_xmind",
    "arguments": {
        "path": "/path/to/output.xmind",
        "sheets": [
            {
                "title": "Overview",
                "rootTopic": {
                    "title": "Project",
                    "linkToTopic": "Details Root",
                    "children": [{"title": "Phase 1"}, {"title": "Phase 2"}]
                }
            },
            {
                "title": "Details",
                "rootTopic": {
                    "title": "Details Root",
                    "linkToTopic": "Project",
                    "notes": {"plain": "Detailed view", "html": "<p><strong>Detailed</strong> view</p>"}
                }
            }
        ]
    }
}

Add the following to your claude_desktop_config.json (on macOS: ~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "xmind": {
      "command": "npx",
      "args": [
        "-y",
        "@41px/mcp-xmind",
        "/path/to/your/xmind/files"
      ]
    }
  }
}

{
  "mcpServers": {
    "xmind": {
      "command": "node",
      "args": [
        "/path/to/mcp-xmind/dist/index.js",
        "/path/to/your/xmind/files"
      ]
    }
  }
}

Restart Claude Desktop after editing the configuration.

A standalone skill is available for creating XMind files without the MCP server. Unlike the MCP server which provides both reading and writing tools, the skill only supports creation.

Build the skill ZIP:

cd skills/xmind && zip -r xmind-skill.zip SKILL.md scripts/

Install: Open Claude Desktop > Settings > Capabilities > Skills > Upload xmind-skill.zip.

The skill script has zero npm dependencies (uses Node.js built-ins only), so it works in the Claude Desktop sandbox.

Capabilities: nested topics, plain & HTML notes (<strong>, <u>, lists), labels, markers, callouts, boundaries, summaries, relationships between topics, internal links across sheets, layout structures (org-chart, fishbone, timeline…), themes (business, dark, simple), simple tasks (todo/done), and planned tasks with Gantt support (dates, duration, progress, priority, dependencies).

MCP server:

claude mcp add xmind -- npx -y @41px/mcp-xmind /path/to/your/xmind/files

Skill (création uniquement): copier ou lier le dossier skill dans .claude/skills/ du projet ou ~/.claude/skills/ pour tous les projets :

# Projet uniquement
ln -s /path/to/mcp-xmind/skills/xmind .claude/skills/xmind

# Tous les projets
ln -s /path/to/mcp-xmind/skills/xmind ~/.claude/skills/xmind

npm run build

npm test

npx @modelcontextprotocol/inspector node dist/index.js /path/to/xmind/files