01What Agent OS Setup Is and How It Works

Agent OS setup is the process of installing and configuring an agent operating system that orchestrates AI agents to automate tasks. In this episode, Julian Goldie explains how the agent OS runs on a local host and how users can customize it to their needs.

The agent OS is a system that lets you build and manage AI agents. It works by letting you create skills that understand your brand, then connecting APIs or MCPs so agents can perform tasks like generating videos from images. The system is designed to be updated daily and comes with full guides and zip files for installation.

02Key Features of Agent OS and Video Tools

The agent OS includes several important features. It is hosted locally, so you have control over your setup. It supports MCPs, which let you plug in tools like Higgsfield for video generation.

You can also integrate it with pay-per-click, an open-source team orchestration system, though Julian notes it can be buggy.

  • Local hosting keeps your agents running on your computer.
  • MCP support allows easy connection to tools like Higgsfield.
  • Custom skills help agents brand videos correctly.
  • Daily updates ensure you have the latest features.
  • Full guides include step-by-step tutorials and zip files.

For video creation, Higgsfield is highlighted as a great tool for generating beautiful videos from images, with an MCP that takes only three steps to set up.

03How to Get Started with Agent OS

To get started with agent OS, you need to install it from the AI Profit Boardroom. Julian provides a zip file and a full step-by-step guide. Here are the basic steps:

  1. Go to the new daily update section on the AI Profit Boardroom.
  2. Find the agent OS system and download the zip file.
  3. Watch the video tutorial for a full setup guide.
  4. Customize it to your needs, or build your own from scratch with the one-hour course.
  5. Connect MCPs like Higgsfield for specific tasks like video generation.

You can also join the community for daily support and coaching calls.

04Using Agent OS: Moving from VPS to Windows

One common question is how to move agent OS from a Linux VPS to Windows. Julian recommends using GitHub to sync the project between environments. This lets you keep your VPS running while developing locally on Windows with tools like Claude Code.

Here is how to do it:

  1. Commit the entire agent OS project to a GitHub repository.
  2. Clone the repository on your Windows local machine.
  3. Continue developing locally, and push changes back to GitHub.
  4. Pull changes on your VPS to keep it in sync.

This method works even if you are using Windows, as GitHub is platform-agnostic. Julian notes this is probably the best solution for moving between environments.

05Restarting Agent OS on Local Host

If your agent OS local host goes down after restarting your computer, you can restart it easily. Julian explains that you simply ask the agent that set up your agent OS to restore and restart the local host.

For example, you can say, "Hey, agent OS isn't open. Can you load it back up for me?" The agent will quickly bring it back online. Sometimes the issue is that a local port isn't being used for the agent OS after restart, but asking the agent usually fixes it.

This is a simple troubleshooting tip that saves time and ensures your agents are always ready to work.

06Using Higgsfield MCP for Image-to-Video

To create videos from images with minimal input, you can use Higgsfield's MCP. This tool lets your agents generate videos without much manual work. Setting it up is simple and takes only three steps.

Once connected, you can plug the MCP into any agent that supports MCPs, such as Hermes, Codex, ChatGPT, or Claude. Here is how to use it:

  1. Create a skill with your AI agent that understands your brand.
  2. Connect the Higgsfield API or MCP to the agent.
  3. Have the agent take an image and generate a video using your skills.

This ensures the videos are branded correctly and match your style.

07Pay-Per-Click Integration and Its Limits

Agent OS includes pay-per-click, an open-source team orchestration system that lets you visualize how agents work together. Julian has it built in for everyone who sets up from his zip file.

However, he admits he doesn't use it much. AI agents are already good at orchestrating themselves, so a full team may not be needed. Pay-per-click can be useful for visual representation, but it can also be buggy sometimes.

Julian finds individual workflows like Hermes Apollo, Hermes Oracle, and Hermes Astros more useful and practical in reality.

08Tips, Limits, and Mistakes to Avoid

To get the most out of agent OS, avoid these common mistakes:

  • Not using GitHub can make moving from VPS to Windows difficult.
  • Ignoring MCPs limits your agents' capabilities.
  • Relying solely on pay-per-click may lead to bugs; use simpler workflows when possible.
  • Not restarting properly after computer shutdown can leave your local host down.

Always ask the agent to restart the local host if it goes down. Also, update your agent OS daily to get the latest features and fixes.

09Who Agent OS Is For and Why It Matters

Agent OS is for anyone who wants to automate tasks with AI agents. It is especially useful for agency owners, like Kevin, who rebuilt his sales funnel in a week. It matters because it lets you create videos, manage workflows, and scale your operations with minimal input.

By using agent OS, you can save time and increase productivity. Julian's community provides support and daily tutorials, so you are never alone. Whether you build your own agent OS or use his, the system is customizable and designed to help you succeed with AI.