MCP Server for Cloudron - AI-Powered Instance Management

[sangemaru]

Hello Cloudron Community!

I'm excited to share mcp-cloudron - an open-source MCP (Model Context Protocol) server that lets AI assistants interact with your Cloudron instance.

What is MCP?

MCP is Anthropic's open standard for connecting AI assistants (like Claude) to external tools and data sources. This server gives AI agents the ability to query and manage your Cloudron instance programmatically.

Current Features (v1.0.0)

• List Applications - See all installed apps with status, health, domain, and resource usage
• Get App Details - Detailed information on specific applications by ID
• Instance Status - Query Cloudron version, configuration, and system info

Quick Start

Install via npm:

npm install @serenichron/mcp-cloudron

Configuration: You'll need a Cloudron API token (Settings → API Tokens). Then add to your MCP client configuration:

{
  "mcpServers": {
    "cloudron": {
      "command": "npx",
      "args": ["-y", "@serenichron/mcp-cloudron"],
      "env": {
        "CLOUDRON_DOMAIN": "your-cloudron.example.com",
        "CLOUDRON_TOKEN": "your-api-token"
      }
    }
  }
}

Use Cases

With this MCP server, AI assistants can:

• Monitor app health and status across your instance
• Get detailed resource usage and configuration info
• Help troubleshoot issues by querying app states
• Provide intelligent suggestions based on your Cloudron setup

Links

• npm Package: https://www.npmjs.com/package/@serenichron/mcp-cloudron
• GitHub Repository: https://github.com/serenichron/mcp-cloudron
• Documentation: Full setup guide and API reference in README

Roadmap

Future features planned for upcoming releases:

• App lifecycle management (start/stop/restart/uninstall)
• App installation from the App Store
• Backup and restore operations
• Log access and monitoring
• Event subscriptions and webhooks

Security Note

The server uses read-only operations in v1.0.0 (list apps, get status). Future write operations will include appropriate safety checks and confirmations. Always use dedicated API tokens with minimal required permissions.

Contributions Welcome!

This project is MIT licensed and I'd love community input:

• Feature Requests: What Cloudron operations would be most useful for AI automation?
• Bug Reports: Issues and feedback on GitHub
• Pull Requests: Code contributions welcome!
• Integration Ideas: How are you using it? Share your use cases!

Testing & Compatibility

• Tested with Cloudron 8.2.0+
• Works with Claude Desktop, Continue.dev, and other MCP-compatible clients
• Node.js 18+ required

Looking forward to your feedback and suggestions!

[james]

Hello @sangemaru and welcome to the Cloudron Community