Coolify MCP Server

Overview

Coolify MCP Server is the leading community-built Model Context Protocol (MCP) server for Coolify, the popular open-source self-hosted PaaS alternative to Heroku, Vercel, and Netlify. It allows AI coding assistants (Claude Desktop, Cursor, Claude Code, etc.) to fully manage and debug your Coolify instance using natural language commands.

With 38 highly optimized tools, smart resource lookup (by name/domain/IP instead of UUIDs), token-efficient responses, and built-in full-text documentation search, it turns your AI into a powerful DevOps co-pilot for self-hosted deployments.

Key Features

• 38 Optimized Tools: Covers infrastructure, servers, projects, applications, databases, services, deployments, environment variables, private keys, teams, cloud tokens (Hetzner, DigitalOcean), and more.
• Smart Diagnostics: diagnose_app, diagnose_server, find_issues with human-friendly identifiers.
• Batch Operations: Restart entire projects, bulk environment updates, stop/redeploy multiple apps.
• Token Efficiency: Responses are 85–99% smaller than raw Coolify API (e.g., list_applications reduced from ~170KB to ~4.4KB).
• Full-Text Docs Search: Built-in search across 1,500+ Coolify documentation chunks.
• HATEOAS-Style Responses: Includes suggested next actions and pagination for better AI context.
• Production-Ready: 98%+ test coverage, strict TypeScript, robust error handling.

Use Cases

• AI-Driven Deployments: "Deploy my Next.js app from GitHub repo X to production."
• Infrastructure Management: "List all servers, show resource usage, and restart the database server."
• Debugging & Diagnostics: "Diagnose why my app is failing and suggest fixes."
• Batch Management: "Update environment variables across all apps in Project Y and redeploy."
• Documentation Assistance: "Search Coolify docs for how to set up PostgreSQL backups."
• Full DevOps Workflows: Combine with other MCP servers for end-to-end AI-orchestrated deployment pipelines.

Installation (Easiest)

Via npx (Recommended for Claude Desktop / Cursor / Claude Code)

# For Claude Desktop – add to claude_desktop_config.json
npx -y @masonator/coolify-mcp

Set these environment variables:

• COOLIFY_ACCESS_TOKEN (your Coolify API token)
• COOLIFY_BASE_URL (e.g. https://your-coolify-instance.com)

Quick one-liner for Claude Code:

claude mcp add coolify -e COOLIFY_ACCESS_TOKEN=xxx -e COOLIFY_BASE_URL=https://your-coolify.com -- npx @masonator/coolify-mcp@latest

Full source and Docker/manual setup available in the repository.

Compatibility

• Clients: Claude Desktop, Cursor, Claude Code, any MCP-compliant AI tool.
• Coolify Version: Tested with v4.0.0-beta.460+ (auto-detects features).
• License: MIT

Community & Alternatives

This is the most actively maintained and feature-rich implementation (v2.7.3 as of February 2026), published to the official MCP Registry. Enhanced forks (e.g., dazeb/coolify-mcp-enhanced) add extra deployment tools if needed.

Start managing your self-hosted PaaS with AI today — Coolify MCP Server makes DevOps conversational and effortless.