dokploy-mcp
Wraps the Dokploy REST API to expose Dokploy operations as MCP tools for managing projects, applications, compose stacks, deployments, domains, and environment variables.
README
@wyattjoh/dokploy-mcp
MCP server that wraps the Dokploy REST API. Exposes Dokploy operations as Model Context Protocol tools for Claude Code, Claude Desktop, and other MCP clients.
Install
npm install -g @wyattjoh/dokploy-mcp
Or with Bun:
bun install -g @wyattjoh/dokploy-mcp
Configuration
The server supports two modes: single-instance and multi-instance. The two modes are mutually exclusive.
Single instance
Set a single Dokploy API token. All tools operate against one Dokploy server.
| Variable | Required | Description |
|---|---|---|
DOKPLOY_API_TOKEN |
Yes | API token for your Dokploy instance |
DOKPLOY_URL |
No | Base URL (defaults to https://dokploy.wyattjoh.dev) |
Multi-instance
Set per-instance tokens and URLs. Tools gain a required instance parameter so
you can target different Dokploy servers from a single MCP session.
For each instance, define a pair of environment variables:
DOKPLOY_<NAME>_API_TOKEN- API token for the instanceDOKPLOY_<NAME>_URL- Base URL for the instance
Instance names are derived by lowercasing the <NAME> segment. Names must match
[a-z0-9]([a-z0-9_]*[a-z0-9])?.
Example for two instances named prod and staging:
DOKPLOY_PROD_API_TOKEN=dkp_abc123
DOKPLOY_PROD_URL=https://dokploy.example.com
DOKPLOY_STAGING_API_TOKEN=dkp_def456
DOKPLOY_STAGING_URL=https://staging-dokploy.example.com
HTTP transport
By default the server uses stdio transport (for Claude Code and similar
clients). Pass --http to start an HTTP server instead.
| Variable | Required | Description |
|---|---|---|
DOKPLOY_MCP_TOKEN |
Yes (HTTP only) | Bearer token for authenticating MCP clients |
dokploy-mcp --http --port 3000
The MCP endpoint is POST /mcp. A health check is available at GET /health.
MCP client configuration
Claude Code (.mcp.json)
Add to your project's .mcp.json or ~/.claude/.mcp.json:
Single instance:
{
"mcpServers": {
"dokploy": {
"command": "dokploy-mcp",
"env": {
"DOKPLOY_API_TOKEN": "dkp_your_token_here",
"DOKPLOY_URL": "https://dokploy.example.com"
}
}
}
}
Multi-instance:
{
"mcpServers": {
"dokploy": {
"command": "dokploy-mcp",
"env": {
"DOKPLOY_PROD_API_TOKEN": "dkp_prod_token",
"DOKPLOY_PROD_URL": "https://dokploy.example.com",
"DOKPLOY_STAGING_API_TOKEN": "dkp_staging_token",
"DOKPLOY_STAGING_URL": "https://staging-dokploy.example.com"
}
}
}
}
Using 1Password CLI (recommended)
Store your Dokploy API tokens in 1Password and use op run to inject them at
runtime. This keeps secrets out of config files entirely.
- Store each API token in 1Password (e.g., in a vault called
Development) - Reference them with
op://URIs in your.mcp.json - Use
opas the command so secrets are resolved in-memory
{
"mcpServers": {
"dokploy": {
"command": "op",
"args": ["run", "--", "dokploy-mcp"],
"env": {
"DOKPLOY_PROD_API_TOKEN": "op://Development/dokploy-prod/credential",
"DOKPLOY_PROD_URL": "https://dokploy.example.com",
"DOKPLOY_STAGING_API_TOKEN": "op://Development/dokploy-staging/credential",
"DOKPLOY_STAGING_URL": "https://staging-dokploy.example.com"
}
}
}
}
With this setup, op run resolves the op:// references and injects the real
tokens into the subprocess environment. No plaintext secrets touch disk.
For HTTP transport with 1Password:
{
"mcpServers": {
"dokploy": {
"command": "op",
"args": ["run", "--", "dokploy-mcp", "--http"],
"env": {
"DOKPLOY_MCP_TOKEN": "op://Development/dokploy-mcp-token/credential",
"DOKPLOY_PROD_API_TOKEN": "op://Development/dokploy-prod/credential",
"DOKPLOY_PROD_URL": "https://dokploy.example.com"
}
}
}
}
Tools
Projects
| Tool | Description |
|---|---|
dokploy_list_instances |
List all configured Dokploy instances |
dokploy_list_projects |
List all projects |
Applications
| Tool | Description |
|---|---|
dokploy_list_applications |
List all applications (filtered from projects) |
dokploy_get_application |
Get full application config by ID |
dokploy_update_application |
Update application config fields |
dokploy_deploy |
Trigger a deployment |
dokploy_start_application |
Start an application |
dokploy_stop_application |
Stop an application |
Compose
| Tool | Description |
|---|---|
dokploy_get_compose |
Get compose stack config by ID |
dokploy_deploy_compose |
Deploy a compose stack |
dokploy_redeploy_compose |
Redeploy a compose stack |
dokploy_start_compose |
Start a compose stack |
dokploy_stop_compose |
Stop a compose stack |
dokploy_update_compose |
Update compose stack config |
dokploy_list_compose_services |
List services in a compose stack |
dokploy_list_compose_deployments |
List deployments for a compose stack |
dokploy_list_compose_domains |
List domains for a compose stack |
Deployments
| Tool | Description |
|---|---|
dokploy_list_deployments |
List deployments for an application |
Domains
| Tool | Description |
|---|---|
dokploy_list_domains |
List domains for an application |
dokploy_create_domain |
Add a domain to an application |
Environment
| Tool | Description |
|---|---|
dokploy_list_environment |
List environment variables for an application |
dokploy_update_environment |
Set environment variables (merge with existing) |
dokploy_delete_environment |
Remove environment variables by key |
Build
| Tool | Description |
|---|---|
dokploy_save_build_type |
Update build configuration |
dokploy_save_github_provider |
Update GitHub source configuration |
Operational
| Tool | Description |
|---|---|
dokploy_get_containers |
List Docker containers |
dokploy_get_app_monitoring |
Get application monitoring data |
dokploy_redeploy_application |
Redeploy an application |
dokploy_cancel_deployment |
Cancel a running deployment |
dokploy_kill_build |
Kill a running build |
dokploy_get_version |
Get the Dokploy server version |
dokploy_list_servers |
List all configured servers |
Development
bun install
bun run start # stdio transport
bun run start:http # HTTP transport on port 3000
bun run build # transpile to dist/ (Node.js target)
bun run lint # oxlint
bun run typecheck # tsc --noEmit
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。