MCP PRODUCT v1.0

> deploio mcp server

A production-ready MCP server for deploying, managing, and monitoring your apps on Deploio infrastructure. Connect any AI client and start Deploying.

mcp-tool-response.json
{
  "tool":   "nctl_apps_list",
  "status": "success",
  "result": {
    "apps": [
      {
        "name":     "my-rails-app",
        "status":   "running",
        "replicas": 3
      },
      {
        "name":     "api-gateway",
        "status":   "deploying",
        "replicas": 1
      }
    ]
  }
}

Everything your AI agent needs

All the Deploio tools your agent needs, surfaced as clean MCP tools — nothing more, nothing less.

🔧

Tool Summary

Automatically introspects available nctl commands and exposes them as typed MCP tools — agents always know what's available.

📶

Streaming Responses

Log output and long-running commands stream token-by-token in real time, so your AI client sees progress as it happens.

Zero Built-ins

No hardcoded command list. Every tool is derived from the live Deploio API surface, so new nctl capabilities appear automatically.

terminal
# 1. Authenticate
$ nctl auth login

# 2. Deploy from git
$ nctl create app myapp \
    --git-url=https://github.com/you/myapp \
    --git-revision=main

# 3. Stream live logs
$ nctl logs app myapp -f

# 4. Run a one-off command
$ nctl exec app myapp -- rails db:migrate

Up and running
in four steps

Connect your AI client to the deploio MCP server and start managing apps with natural language.

  1. Install & authenticate nctl

    Install the Deploio CLI and run nctl auth login to connect your account.

  2. Start the MCP server

    Run the MCP server locally or via Docker. It auto-discovers all available nctl tools on startup.

  3. Connect your AI client

    Point Claude, Cursor, or any MCP-compatible client at the server endpoint.

  4. Deploy with natural language

    Ask your agent to deploy, scale, or inspect apps. It calls the right Deploio tools automatically.