1. Conduid
  2. Analytics
  3. Daemonize
MCP server · Analytics

Daemonize

A Model Context Protocol server for seamless management and monitoring of long-running daemons via AI agents

Unclaimed mcpmcp-server
34Low

Scored 6 months ago · breakdown

About Daemonize

Daemonize is an MCP server in the Analytics category: a Model Context Protocol server for seamless management and monitoring of long-running daemons via AI agents.

Install

From source
git clone https://github.com/mackee/mcp-daemonize

This server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.

Ask AI

✦

Ask AI about Daemonize

Powered by Claude · Grounded in docs

I know everything about Daemonize. Ask me about installation, configuration, usage, or troubleshooting.

Security checks

  • ✓README presentThe repository ships documentation.

5 checks have not been run yet.

README

mcp-daemonize

mcp-daemonize is a Model Context Protocol (MCP) server designed to provide seamless management of long-running daemons (such as development servers) for AI agents like Claude Code and Cline. It enables advanced automation and debugging capabilities by allowing AI agents to start, stop, and monitor daemons, as well as access their logs in real time.

Motivation

While AI agents can execute commands, they are not well-suited for managing long-running development servers (e.g., vite dev). This is because the agent typically waits for the command to finish, which is not suitable for servers that are meant to run continuously during development. As a workaround, users often start development servers in a separate shell, but this prevents the AI agent from directly accessing the server's logs, making autonomous debugging impossible.

mcp-daemonize solves these problems by providing:

  • An interface to start daemons without waiting for their termination.
  • Tools to stop running daemons.
  • Tools to view real-time logs of running daemons.

With these features, mcp-daemonize enables AI agents to fully manage development servers, monitor their output, and perform automated debugging, greatly expanding the possibilities for autonomous coding.

Use Cases

  • Starting and stopping development servers (e.g., Vite, Next.js, etc.) from an AI agent.
  • Viewing and analyzing real-time logs of running daemons for debugging.
  • Automating development workflows that require persistent background processes.
  • Enabling AI agents to manage, monitor, and debug long-running processes autonomously.

Prerequisites

  1. Go 1.24.2 or later installed.
  2. (Optional) Docker, if you wish to run mcp-daemonize in a container.

Installation

Download from GitHub Releases

Visit the Releases page to download the latest version.

  1. Download the archive for your platform from the latest release page.
  2. Extract the archive and move the binary to a directory in your PATH.
wget https://github.com/mackee/mcp-daemonize/releases/download/vX.Y.Z/mcp-daemonize_X.Y.Z_Darwin_x86_64.tar.gz
tar -xzf mcp-daemonize_X.Y.Z_Darwin_x86_64.tar.gz
sudo mv mcp-daemonize /usr/local/bin/

Build from Source

You can install the latest version using go install (Go 1.24.2 or later):

go install github.com/mackee/mcp-daemonize/cmd/mcp-daemonize@latest

Usage with MCP Host (e.g., Claude Code, Cline)

Add the following configuration to your MCP host settings (e.g., mcp.json):

{
  "servers": {
    "daemonize": {
      "command": "/path/to/mcp-daemonize",
      "args": [],
      "env": {}
    }
  }
}

Replace /path/to/mcp-daemonize with the actual path to the built binary.

Usage

mcp-daemonize provides the following tools for AI agents:

Tools

  • daemonize_start

    • Start a long-running process (e.g., a development server) as a daemon.
    • Parameters:
      • name (string, required): Name of the daemon.
      • command (string[], optional): Command to run (e.g., ["npm", "run", "dev"]).
      • workdir (string, required): Working directory for the daemon (absolute path).
  • daemonize_stop

    • Stop a running daemon by name.
    • Parameters:
      • name (string, required): Name of the daemon to stop.
  • daemonize_list

    • List all currently running daemons.
    • Parameters: None
  • daemonize_logs

    • Retrieve the latest logs from a running daemon.
    • Parameters:
      • name (string, required): Name of the daemon.
      • tail (number, required): Number of lines to read from the end of the log.

Example Workflow

  1. Start a development server as a daemon using daemonize_start.
  2. Use daemonize_logs to monitor the server's output in real time.
  3. Stop the server with daemonize_stop when finished.

License

This project is licensed under the terms of the MIT open source license. See LICENSE for details.

README copy captured when this server was indexed; it may lag the repository. The original is authoritative.

Questions

About Daemonize

How do I install Daemonize?

No package has been published for this server, so install it from source: clone the repository and follow its README.

Is Daemonize safe to use with an AI agent?

Its trust score is 34 out of 100 (low). It passes 1 of the 1 static security checks run so far. It has no ConduID identity yet, so agent calls to it are not receipted.

Is Daemonize still maintained?

Conduid hasn't recorded a commit date for this repository yet. Check the repository directly for recent activity.