An AI agent operating system is a single workspace that hosts multiple AI agents and their workflows side by side. The setup demoed by SEO marketer Julian Goldie in October 2026 combines Claude Code, Antigravity and Hermes modules for video, search and sales work, built iteratively over months rather than shipped as a product.
What an AI Agent Operating System Actually Is
An AI agent operating system is a single workspace that runs multiple AI agents and their workflows in one interface, instead of one chat window per tool. The version shown in the video by Julian Goldie, an SEO marketer, bundles agents like Claude, Anthropic AI assistant, with modules for video production, keyword research, content deployment and sales analysis. He presents it as a fix for tool sprawl: too many subscriptions, too many apps, outputs lost in local folders.
The important context is how it was made. Goldie states in the video (published 2026-10-09) that he spends roughly three to four hours a day building and improving the system, and that it did not look like this on day one. It is a personal, iteratively grown setup, and he distributes it through his AI Profit Boardroom community as a zip file with a 30-day installation roadmap, a paid offer that shapes the video's tone.
Which Agents Plug Into the Workspace
The workspace is model-agnostic by design: the presenter says you can plug in whichever agents you want and change the model at any time. The video names four specific agents. Claude handles general assistance, Claude Code, Anthropic agentic coding tool that runs in the terminal, is referenced for building inside the system, OpenAI Codex, OpenAI's coding agent, appears in the 3D worlds module, and Antigravity, Google's agentic development platform, is listed as another pluggable agent.
The presenter also builds custom workflows on top of Hermes, which he describes as an agent most people only use inside a plain chat interface. Two named Hermes workflows are Hermes Muse, which analyzes his recent content and suggests ideas, and Hermes Astros, which reviews competitor content and proposes new titles and angles that feed his SEO content agent.
| Agent | Vendor | Role in the demo | Link verified |
|---|---|---|---|
| Claude | Anthropic | General assistant inside the workspace | Yes |
| Claude Code | Anthropic | Terminal coding agent referenced for building | Yes |
| Codex | OpenAI | Powers the 3D worlds module with Blender | Yes |
| Antigravity | Pluggable agentic development platform | Yes |
What the Modules Do: Video, Search, Sales
Higgsfield media module
A section built around Higgsfield, an AI video-generation platform, produces short films, UGC-style ads, cinematic ads and explainer videos. The presenter also regenerates images from the same panel, so an ad created two weeks earlier can be pulled up and remade with one click.
Search OS
The Search OS module replaces, in his words, a collection of five or six separate tools. It runs keyword research across his sites, generates content from a case study plus target keywords, and deploys pages to his websites. This is the presenter's description of his own setup, not an independent review of results.
Sales agent OS
A sales module analyzes recorded sales calls and scores each rep against the team's standard operating procedure, comparing individual performance with the rest of the team. Rep identities are anonymized in the demo.
Video and 3D workflows
A video module writes the script, generates the video, and stores the finished file inside the workspace. A separate 3D worlds module, combining Codex and Blender, the open-source 3D creation suite, lets the presenter walk through and build 3D environments from the same interface.
Why One Workspace Beats Scattered Tools
The core argument is retrieval and continuity. The presenter says a common failure mode is building something useful, then losing it in a local folder or inside a subscription you later cancel. Central storage means outputs stay addressable: the ad from two weeks ago is one click away, and regeneration is built into the panel.
The second argument is composability. Because the modules share a workspace, an idea generated by the competitor-analysis workflow can be plugged straight into the SEO content agent or the video agent. He adds custom pieces as needed, such as a voice agent and an 'AI agent mastermind' mode where agents collaborate on tasks, and he can hide modules he stops using, like his lead-generation tool, with a checkbox.
How the System Was Built: Iteration Over Weeks
The build sequence described in the video is weekly and incremental. In the most recent week he added a pixel-art office view where agents appear as characters at desks; the week before he created the Search OS UI; before that he improved deployment systems and built more websites; and the week before that he set up the Higgsfield MCP integration. Each step addressed one workflow.
His advice for anyone copying the pattern: do not build everything at once. He compares the process to building a SaaS product, where pruning unused features is normal. Automating one thing usually reveals the next thing worth automating, so the system grows as a recurring weekly habit rather than a single project.
What to Weigh Before Copying This Setup
Every claim in the demo comes from the person selling the setup. The video doubles as marketing for the AI Profit Boardroom community, which sells the zip file, roadmap, courses and coaching calls, so statements like 'so much time saved' are the presenter's first-hand experience, not a measured benchmark. No independent testing of the workspace appears in the video.
The agent names also carry fine print. Codex, Claude Code and Antigravity are coding and development agents; their effectiveness inside a custom UI depends on the workflow built around them, which this presenter spent months tuning. If you replicate the pattern, expect to hide modules, rebuild parts, and accept that the polished demo reflects roughly three to four hours of daily work by his own account.
Frequently Asked Questions
- What is an AI agent operating system? It is a single workspace that hosts multiple AI agents and their workflows together, so video generation, keyword research and other tasks run from one interface instead of separate tools and chat windows.
- Which AI agents does the demoed system use? The video names Claude, Hermes, Antigravity and Codex as pluggable agents, and the presenter says the model can be changed at any time. Claude Code is also referenced for building inside the system.
- Is the setup ready to use out of the box? No. The presenter states he spent about three to four hours a day building it and recommends installing it step by step using the 30-day roadmap he sells through his community.
- Was the system independently tested? No. All performance and time-saving claims in the video come from its creator, who also sells the setup, so treat them as first-hand experience rather than verified results.
Fork this article
Start a new branch from the same video, shaped your way. You keep the credit; the original keeps the attribution.
A fork in another language is filed as a translation of this article, so the two pages point at each other. You can unlink it later from the editor.
0/240
You are creating
- Format
- For
- Language
- Source
- Your angle
You will be asked to sign in before it is generated.
Buy credits