01What an AI agent OS is and why Julian built one

An AI agent OS is a single system that runs all your AI agents in one place. Julian Goldie's version plugs in Claude, Hermes, Codex and Anti-Gravity, so you use one hub and one workspace.

You do not jump between tabs or copy and paste between tools. Everything lives in one place.

Julian built it because of what he sees as the number one problem with AI right now: it is overwhelming. There are too many tools, too many subscriptions and too many workflows scattered across too many apps. People spend more time managing their tools than using them.

The AI agent OS solves that by putting the tools together. In this episode Julian walks through what is inside:

  • Search
  • Content
  • Video
  • Voice agent
  • Mastermind
  • 3D worlds
  • The agent office
  • Open source video

02Search and content inside the AI agent OS

The first sections Julian shows are search and content.

Search. Normally, keyword research for the AI Profit Boardroom means opening four or five different tools, running separate searches, exporting data and importing it somewhere else. In the AI agent OS you go to the search section and run it. It is done, and everything stays inside the system, so nothing gets lost.

Content. Say you want a new article for the AI Profit Boardroom that targets a specific keyword and is built around a case study. The steps are short:

  1. Open the content section.
  2. Plug in the keyword.
  3. Plug in the case study.
  4. Hit go.

The agent writes the article and deploys it.

The point is that research and publishing sit in the same place. You do not export a list from one tool and import it into another before writing.

03Video and storage in the AI agent OS

The video section works from a short description. You type in what the video is about. Julian's example is a short explainer on why AI automation matters for business owners joining the AI Profit Boardroom.

The system writes the script, creates the video and drops it straight into your workspace. It sits there, ready whenever you want to come back to it.

Julian says people underestimate the next part. Everything you create is stored inside the system. One of the biggest hidden problems with AI tools is that people build something great and then can never find it again.

It is in a local folder, or inside a tool they stopped opening, or they cancelled the subscription and it is gone.

In an AI agent OS you get:

  • One place for scripts, videos and articles
  • Work you can find again later
  • No dependence on one tool's subscription

04Voice agent and mastermind in the AI agent OS

Two sections change how you work with agents.

Voice agent. Instead of typing every request, you talk. Say you are working on a new onboarding flow for the AI Profit Boardroom and want to change how something works. You say it out loud and it updates in real time.

Julian built this directly into the system, and it changes how fast you can move.

Mastermind. This is where multiple agents work together at the same time. Instead of asking Claude something and then going to Hermes separately, you have both running, talking to each other and building together.

For example, you say: come up with a new content angle for the AI Profit Boardroom. Both agents start. One pulls the research and one generates the ideas.

You get everything in one place.

This is a different way of using AI. Instead of picking up one tool, putting it down and picking up another, your agents work as a team.

05The 3D worlds, agent office and open source video

Julian also shows three sections that go beyond text.

  • 3D worlds. He built full 3D worlds using Codex and Blender inside the OS. You can walk in, walk out, build environments, save them and come back to them. It shows the point of the OS: you build something and it is right there every time you return.
  • The agent office. You see agents at work: a web designer agent at one desk, a marketer at another and a tester engineer in the background. You walk around, monitor them, create a new agent when you need one and plug in whatever CLI you want. Julian says part of it is fun, but the real value is visibility. You see what is running, what is done and what needs attention.
  • Open source video. Open Montage creates cinematic style videos that look good, not standard AI video output.

There is also a call analysis section that checks each call against your SOP and shows where each rep performs well and where they lose ground.

06Tips for building your own AI agent OS

Julian is honest that the system did not look like this from day one. He spends around 3 to 4 hours a day improving and adding to it, and something new arrives every week.

He lists a few recent additions:

  • The search OS
  • Better deployment
  • Starnet
  • The Higgsfield MCP connection

His lesson is that every time you automate one thing, you notice the next thing that needs automating. The people who build the best AI setups treat it like a product they keep improving, not something they set up once and walk away from.

Two more tips from the episode:

  1. You do not need all of it. If you do not use a section, such as the lead generation module, tick the box and it is hidden.
  2. Let the system shape around your workflow, not the other way around.

07Who the AI agent OS is for and how to get it

An AI agent OS is for business owners who use several AI tools and are tired of managing them. Julian says most people are still jumping between five different AI tools with no system holding it together. The people building unified setups now will be well ahead by the time everyone else catches up.

Inside the AI Profit Boardroom you can download the full Agent OS as a zip file, the same system shown in the video. There is a 30 day roadmap:

  1. Day one: install the system and connect your first agent.
  2. Day two: build your search OS workflow.
  3. Day three: set up your content agent and run your first automated deployment.

By the end of week one you have a working multi-agent hub. Members also get four coaching calls a week, daily tutorials, a prompt library and a member map.

To start for free, Julian points to the AI Success Lab, with the video notes and over 100 AI use cases.