Coolify AnythingLLM

An official Model Context Protocol (MCP) server that connects your AnythingLLM chat interface directly to your self-hosted Coolify infrastructure. Manage, deploy, and troubleshoot your production servers using natural language.

Developed and maintained by the team behind xlop.cz – the Czech tech hub for AI & Self-Hosting innovations.

---

🧠 Why this project?

Giving AI agents the power to write code is great, but giving them the context of your infrastructure is a game-changer. This MCP server acts as a secure bridge:

1. AnythingLLM acts as the intelligent agent and UI.
2. MCP (Model Context Protocol) provides the secure communication standard.
3. Coolify API executes the infrastructure tasks on your home lab or VPS.

🌟 Key Features

• App Diagnostics: Ask your AI about active applications, container status, and resources.
• Log Fetching: Let the AI read application and deployment logs to troubleshoot bugs in seconds.
• Deployment Automation: Trigger new builds, restarts, or rollbacks using simple prompts.
• Secure by Design: No hardcoded credentials. Everything runs via Environment Variables.

🛠️ Installation & Setup

1. Prerequisites

• A running Coolify instance with an API Token (generated in your Coolify settings).
• AnythingLLM (Desktop app or self-hosted via Docker).
• Node.js (v18+) or Docker environment to run this MCP server.

2. Environment Setup

Create a .env file in the root directory of this project:

COOLIFY_API_URL=[https://your-coolify-instance.com/api/v1](https://your-coolify-instance.com/api/v1)
COOLIFY_API_TOKEN=your_secret_coolify_api_token

3. Connect to AnythingLLM

Open your AnythingLLM settings, navigate to Agent Custom Tools / MCP Servers, and add this server using the following configuration:

• Mode: stdio
• Command: node
• Args: ["/path/to/coolify-anythingllm-mcp/dist/index.js"]

(Alternatively, run it via Docker using the provided Dockerfile).

🤖 Awesome Prompts to Try

Once connected, open a chat in AnythingLLM and try these:

• "List all running applications on my Coolify server."
• "Check the deployment logs for my frontend app. Why did it fail?"
• "Restart the PostgreSQL database container immediately."

⚠️ Disclaimer & Safety

Running AI with infrastructure access is powerful but experimental.

• We strongly recommend running your AI agent with Human-in-the-loop confirmations enabled.
• The authors are not responsible for any server downtime, data loss, or accidental resource deletion caused by your AI model.

📄 License

This project is open-source and licensed under the MIT License - see the LICENSE file for details.

---

Built with ❤️ for the open-source community by xlop.cz.