Coolify MCP Server
An MCP server that enables AI assistants to manage infrastructure, applications, databases, and services through the Coolify PaaS platform.
README
Coolify MCP Server
<img src="graphics/CoolifyMCP.png" width="256" alt="Coolify MCP Logo" />
A comprehensive Model Context Protocol (MCP) server for the Coolify API. This server enables AI assistants to manage infrastructure, applications, databases, services, and more through your self-hosted Coolify PaaS platform.
Compatibility: Tested with Coolify version 4.0.0-beta.418
👋 Open for Opportunities!
I'm actively seeking new opportunities and collaborations. If you have interesting projects or job opportunities, feel free to reach out!
📧 Contact: pashvc@gmail.com
Prerequisites
- Node.js 18 or higher
- Coolify Base URL
- Coolify API token
Features
Core Management
- Teams: List teams, get team details, view team members
- Projects: Create, list, update, and delete projects
- Servers: Create servers, validate connections, view resources and domains
- Private Keys: Manage SSH keys for server access
Applications
- Full CRUD operations (create, read, update, delete)
- Lifecycle management (start, stop, restart)
- Execute commands in containers
- View application logs
- Git integration (repository, branch, commit management)
- Resource limits configuration (memory, CPU)
- Health check configuration
Databases
- Support for PostgreSQL, MySQL, MariaDB, MongoDB, and Redis
- Create and manage database instances
- Configure database-specific settings
- Lifecycle management (start, stop, restart)
- Resource limits and public port configuration
Services
- Deploy pre-configured services (Plausible, Umami, n8n, etc.)
- Service lifecycle management
- Configuration updates
Environment Variables
- Manage environment variables for applications, services, and databases
- Support for build-time and runtime variables
- Preview deployment variables
- Shared variables across environments
Deployments
- List all deployments
- Get deployment details and logs
- Trigger deployments via webhook
- Support for Docker image tags
Developer Experience
- Pre-configured prompts for common workflows
- Comprehensive error handling with detailed messages
- TypeScript support with full type safety
- Integration with MCP Inspector for debugging
- Robust API response handling for different Coolify versions
- Extensive test coverage with both unit and integration tests
Installation
# Install globally
npm install -g @pashvc/mcp-server-coolify
# Or use with npx
npx @pashvc/mcp-server-coolify
Configuration
The server requires two environment variables:
COOLIFY_BASE_URL: The base URL of your Coolify instanceCOOLIFY_TOKEN: Your Coolify API token
Getting an API Token
- Go to your Coolify instance
- Navigate to
Keys & Tokens/API tokens - Create a new token with the following required permissions:
- read (for fetching information)
- write (for managing resources)
- deploy (for deployment operations)
Usage
In MCP Settings
Add the following to your MCP settings configuration:
{
"mcpServers": {
"coolify": {
"command": "npx",
"args": ["-y", "@pashvc/mcp-server-coolify"],
"env": {
"COOLIFY_BASE_URL": "your-coolify-url",
"COOLIFY_TOKEN": "your-api-token"
},
"disabled": false,
"autoApprove": []
}
}
}
Windows Cline users may need the following:
{
"mcpServers": {
"coolify": {
"command": "cmd",
"args": [
"/c",
"npx",
"-y",
"@pashvc/mcp-server-coolify"
],
"env": {
"COOLIFY_BASE_URL": "your-coolify-url",
"COOLIFY_TOKEN": "your-api-token"
},
"disabled": false,
"autoApprove": []
}
}
}
Available Tools
Version & Health
coolify_version: Get Coolify version informationcoolify_health: Check Coolify API health status
Teams
list_teams: List all teamsget_team: Get details of a specific teamget_current_team: Get current team detailsget_current_team_members: Get current team members
Servers
list_servers: List all serverscreate_server: Create a new servervalidate_server: Validate server configurationget_server_resources: Get server resourcesget_server_domains: Get server domains
Projects
list_projects: List all projectsget_project: Get project detailscreate_project: Create a new projectupdate_project: Update project settingsdelete_project: Delete a project
Applications
list_applications: List all applicationsget_application: Get application detailscreate_application: Create a new applicationupdate_application: Update application settingsdelete_application: Delete an applicationstart_application: Start an applicationstop_application: Stop an applicationrestart_application: Restart an applicationexecute_command: Execute command in application containerget_application_logs: Get application logs
Databases
list_databases: List all databasesget_database: Get database detailscreate_database: Create a new databaseupdate_database: Update database settingsdelete_database: Delete a databasestart_database: Start a databasestop_database: Stop a databaserestart_database: Restart a database
Services
list_services: List all servicesget_service: Get service detailscreate_service: Create a new serviceupdate_service: Update service settingsdelete_service: Delete a servicestart_service: Start a servicestop_service: Stop a servicerestart_service: Restart a service
Environment Variables
list_environment_variables: List environment variablescreate_environment_variable: Create a new environment variableupdate_environment_variable: Update an environment variabledelete_environment_variable: Delete an environment variable
Deployments
list_deployments: List all deploymentsget_deployment: Get deployment detailsdeploy_webhook: Trigger deployment via webhook
Private Keys
list_private_keys: List all private keyscreate_private_key: Create a new private keyupdate_private_key: Update a private keydelete_private_key: Delete a private key
Pre-configured Prompts
The server includes helpful prompts for common workflows:
deploy-git-app: Deploy a new application from a Git repositorysetup-database: Set up a new database with proper configurationdeploy-service: Deploy a pre-configured serviceenvironment-setup: Set up environment variables for an applicationserver-health-check: Comprehensive health check of all serversbackup-database: Create a backup of a databasescale-application: Scale an application by adjusting resourcessetup-ssl: Configure SSL/TLS for an applicationmigrate-application: Migrate an application between serverstroubleshoot-deployment: Troubleshoot a failed deployment
Development
Building from Source
# Clone the repository
git clone https://github.com/pashvc/mcp-server-coolify.git
cd mcp-server-coolify
# Install dependencies
npm install
# Build the project
npm run build
# Run tests
npm test
Development Scripts
npm run build: Build the TypeScript projectnpm run watch: Build and watch for changesnpm test: Run all testsnpm run test:watch: Run tests in watch modenpm run test:coverage: Run tests with coverage reportnpm run test:integration: Run integration tests only (requires env vars)npm run inspector: Launch MCP Inspector for debugging
Troubleshooting
Common Issues
- 404 Errors: Some endpoints may not be available in all Coolify versions. The server automatically handles different response formats across versions.
- Authentication Errors: Ensure your API token has the required permissions (read, write, deploy).
- Connection Issues: Verify your
COOLIFY_BASE_URLis correct and accessible. - Environment Variables: Make sure both
COOLIFY_BASE_URLandCOOLIFY_TOKENare set correctly.
Testing Your Setup
# Test with your credentials
export COOLIFY_BASE_URL="https://your-coolify-instance.com"
export COOLIFY_TOKEN="your-api-token"
# Run integration tests to verify connectivity
npm run test:integration
# Or test the server directly
npm run build && npm start
API Compatibility
The server handles different response formats automatically:
- Version endpoint: Supports both string and object responses
- List endpoints: Handles both direct arrays and paginated responses
- Health endpoint: Adapts to different status response formats
- Error handling: Robust parsing of various error message formats
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。