01What Is an Agent Operating System and How It Works

The agent operating system is a powerful tool that controls all your AI agents, orchestrates tasks, and automates workflows. It gets updated daily, so you always have the latest features.

Think of it as a central hub where you can build, customize, and run agents for many purposes—like lead generation, outreach, voice control, and even game creation. It works by letting you create separate workflows for different jobs, and you can connect it to local models or cloud APIs.

In this episode, Julian Goldie shows how the system can be used on multiple machines, with a lead machine built in, and even with free AI models. The system is yours to keep, unlike models that can be taken down anytime.

02Key Features of the Agent OS

The agent operating system comes with many built-in features that make it powerful and flexible. Here are the main ones shown in this episode:

  • Voice agent – You can talk to it, use a wake word, and it responds in real time.
  • Lead machine – Generates validated leads, enriches them with company details, and runs email campaigns.
  • Fusion – A feature that combines multiple agents to get better outputs.
  • Memory galaxy – Visualizes your memories in a galaxy-like view, and lets you plug in contacts instantly.
  • Local agent control – Run agents locally on your machine for more security and control.
  • Custom workflows – Create sections for specific tasks like SEO, video, or outreach.

You can also use free models like North Mini from OpenRouter or Olama, so you don’t have to pay for APIs.

03How to Install and Set Up the Agent OS

Getting started with the agent operating system is straightforward, but there are a few key steps. Julian recommends installing everything in one place locally to keep things simple.

  1. Get the agent OS from the classroom – go to the daily updates section and find the Agent OS.
  2. Download the zip file and follow the video tutorial and step-by-step guide.
  3. Install it on your main machine – or, if you use multiple machines, use TailScale to link them together.
  4. For a messy setup, set up a VPS and secure it with Cloudflare Access or TailScale.
  5. Once installed, start creating workflows for different tasks like SEO, video, or outreach.

If you run multiple businesses, you can use the same system by creating separate workflows for each one. That way, everything stays organized.

04How to Use the Agent OS for Lead Generation

One of the most powerful uses of the agent operating system is turning it into a lead machine. Here’s how it works, as shown in this episode:

  1. Go to the outreach section and type in the type of leads you want.
  2. The system starts generating a list with domain, company, status, and notes.
  3. It validates emails to make sure they are real and reachable.
  4. Delete any leads you don’t want from the list.
  5. Go to the campaigns tab and create a new campaign.
  6. Tell the AI what the campaign is about, or write the subject and email manually.
  7. Preview the email and click “create campaign.”

Once live, you can see how many leads you’ve contacted, how many are sendable, and it automatically removes companies that are too big to reach out to. This is a full SaaS tool built into the agent OS.

05Testing Results and Community Examples

In this episode, Julian shows real examples from the community. Sheena turned her agent OS into a self-running lead machine that looks like a full website. It’s an interactive demo of her AI agents, and it generates leads for her agency.

Julian also shows Robert’s success story – he started as a CRO four years ago and is now crushing it with the system. And there’s the 3JS game skill that improved game quality when installed into Claude.

Julian tested HyperFrames and ReMotion for agent pipelines. He found HyperFrames (HTML) more reliable and better-looking, while ReMotion (React and TypeScript) was slower. For avatars, they default to Haygen.

These examples show that the system works for different use cases – from lead generation to games – and you can customize it to fit your needs.

06Tips, Limits, and Mistakes to Avoid

There are some important tips and warnings from this episode. First, avoid fragmenting your setup. Install everything in one place locally to make it easier to manage.

If you must use multiple machines, use TailScale or Cloudflare Access to link them securely.

For lead generation, the system skips companies that are too big to reach out to, but you can keep them if you want. That’s a smart feature to avoid wasting time.

When using free models, you can get good outputs, but be careful – models can be gated or removed anytime. That’s why Julian focuses on the system, not the model. Use the system with free models if you want, but know the risks.

For security, Julian advises caution with open-source tools like OpenClaw. If you’re not sure about security, don’t use it. Stick with stable CLI options like Claude CLI or GLM 5.2 CLI instead.

07Who Should Use the Agent OS and Why

The agent operating system is for anyone who wants to automate tasks, generate leads, or build AI-powered workflows. It’s especially useful for business owners, marketers, and developers.

You can run multiple businesses in one system, create custom workflows, and even use free models to keep costs low. Julian says the system is yours, so you can hold on to it and automate anything you want – no matter what happens with models.

It matters because models are becoming less available – like GPT 5.6 being in preview – but systems give you control. With consistent updates and a community of 3,800 members, you can get help and learn from others.

Whether you’re a beginner or advanced, you can start with the daily updates, join coaching calls, or meet people locally. The agent OS helps you win with AI.