AI OS FAQs

An AI OS is an AI Operating System: the layer underneath your AI tools that holds your context, your voice, your offers, your customers and your workflows in one structured place every AI you use can read.

Think of how macOS or Windows loads your files, your apps and your preferences the moment you log in. You do not tell it who you are every morning. It already knows.

An AI Operating System does the same thing for an AI session. Without one, each tool starts from zero every time you open it and you re-explain your business forever. With one, the work compounds.

Your AI forgets because a chat session has no memory of your business unless something loads it at the start.

Every tool you open begins with a blank context window. It knows the model it was trained on and nothing about you. So you paste in the same background, correct the same mistakes, and explain your offer again for the hundredth time.

An AI OS fixes the cause rather than the symptom. Your business lives in plain files in your AI OS workspace folder, and the agent reads them before you type anything.

An AI OS holds 5 things, and every one of them is a plain text file you own.

• Context. Who you are, what your company does, who buys from you, and the tools you already run.
• Voice. How you write, and the words you never use.
• Rules. What the agent must do and must never do, read first in every session.
• Memory. What you decided, what you corrected, and what came up while working.
• Skills. The jobs it already knows how to do, invoked by name.

The free kit ships 6 context files and 12 skills. Every word in them is yours to read and change.

A prompt library is a drawer of instructions you go and fetch. An AI OS is loaded before you ask for anything.

With a library you still have to remember which prompt to use and paste your context in every time. With an AI OS your voice, your offers and your customers are already in the session, so the agent starts from your business instead of from nothing.

Prompts sit on top of an AI OS. They do not replace it.

A second brain is written for you to read. An AI OS is written for an agent to read and act on.

Notes sit there until you open them. Your AI OS gets loaded at the start of every session, and the skills and routines on top of it do work rather than store it.

Point Obsidian at the same folder if you like reading it that way. Nothing changes.

A custom GPT lives inside 1 vendor's product. An AI OS lives in your AI OS workspace folder, which every vendor can read.

That difference decides what happens next year. A custom assistant is a copy of your context stuck in someone else's account, and it does not come with you when a better model ships.

Your AI OS is plain markdown you own. When the next agent arrives you point it at the same folder, and everything you have written keeps working.

Yes, and daily users usually feel the difference fastest.

If you are prompting well but starting from scratch in every session, the system underneath the prompts is the part you are missing. Better prompting does not fix a blank context window.

Most people who already use AI daily find the compounding starts immediately, because their context is finally written down somewhere the agent reads on its own.

The AI OS Starter Kit is a free guide and workspace that takes you from an empty folder to a working AI Operating System.

You copy the folder to your own computer, run 1 command, and answer some questions. From then on, every session starts with your agent already knowing your business, your customers, your voice and your tools.

It is the free entry point to the AI OS Navid runs his own business on, with his specifics taken out and a setup interview put in.

Yes. The kit is MIT licensed and you get all of it.

There is no credit card, no trial period and no locked section you pay to open. You can read every file, change any of it, and use it in your own business with no strings.

There is no catch. The kit is the free entry point to the way Navid works.

The people who get value from it are the ones who later come to the AI OS Workshop to build the full system live, or buy AI OS Pro. You are not required to buy anything, and the kit is not crippled to push you into it.

It is the real workspace. Plenty of people run it for months without spending a cent.

The starter kit is the workspace itself, set up and ready for your answers.

• 6 context files holding who you are, your company, your voice, the words you never use, your design taste and your tools.
• 12 slash commands, from a morning brief to meeting notes, research, decks, diagrams and the commands that keep the OS current.
• A CLAUDE.md that tells the agent how to work before you type anything, with AGENTS.md and GEMINI.md pointing at the same file.
• A brand-guardian agent that checks every draft against your voice and rejects the ones that miss.
• Folders for outputs, projects, daily notes, decisions and open work.

All of it is plain markdown. Read it in any editor.

The starter kit is for creators, founders and operators who already use AI and can feel they are getting a fraction of what it should give them.

You need a laptop, an agent you already use, and about 15 minutes. You do not need to know how to code.

It is not for you if you want a chatbot, or if you are not willing to write your business down. The setup does most of that work and it still needs your answers.

Yes. The kit is on GitHub at thenavidm/ai-os under the MIT licence.

That means you can read every line, fork it, change it and use it commercially. It is also published to npm as @thenavidm/ai-os.

AI OS Pro is a separate repository and is licensed rather than open source.

The email course is 7 lessons on how Navid actually runs his AI OS, sent 1 a day.

Lesson 1 arrives as soon as you sign up, with the starter kit itself. The 6 after it go deeper into the parts people get stuck on.

You can take the kit and ignore the emails. Every one of them has a 1-click unsubscribe that works immediately.

You install it with 1 command that copies the folder and opens your agent inside it.

The command is on the home page and in the repository readme. It copies the folder without any git history, so the OS is yours from the first commit rather than a fork of anyone else's. It then checks that Claude Code is installed, makes the folder a git repo of your own, and opens the agent in it.

Prefer to do it by hand? Download the ZIP from GitHub, unpack it, and open the folder in your agent the way you normally do. Your agent reads CLAUDE.md from the folder it is opened in. That is how the OS loads.

The install is 1 command. The setup interview takes about 15 minutes.

That interview is the part that matters. It connects what it can reach, reads your own sent mail and posts before it asks anything, then asks 12 questions one at a time with examples and previews every file before it writes.

Connecting your tools adds a few minutes per service, and you do those whenever you want rather than all at once.

The setup interview asks 12 questions about your business, one at a time, each with an example so you are never staring at a blank prompt.

It covers who you are, what you sell, who buys from you, how you sound in writing, the words you refuse to use, and the tools you already run. It reads your real writing first, so several answers arrive already drafted for you to correct rather than compose.

Every answer is previewed before anything is written to a file.

Yes. Stop whenever you like.

Every answer is saved the moment it lands, and running the setup command again continues from where you stopped. Nothing is lost and nothing is asked twice.

That is deliberate. Writing your business down honestly is worth doing properly, and forcing it into 1 sitting is how people write filler instead.

No. The agent handles the technical work.

Your job is to describe what you want and review what it builds. The kit is written for non-technical creators and founders and it walks through every step.

If you can edit a text file and follow a numbered list, you can run an AI OS.

You need a laptop, an AI agent you already use, and a text editor.

• An agent. Claude Code, Codex, Cursor or Gemini CLI all read the same folder.
• Node 18 or newer, which runs the install script.
• A text editor. Any one works.

Claude Code is included with the Claude Pro plan, which is enough for almost everyone starting out. You do not need the Max plan.

Yes. Obsidian is the nicest way to read and edit the files yourself.

Download Obsidian, choose Open folder as vault, and point it at your AI OS folder. Everything is markdown, so it renders properly and the internal links work.

Your agent does not need it. This is purely for you.

An AI OS works with any agent that can read a folder, which today means Claude Code, Codex, Cursor and Gemini CLI.

The guide is built around Claude Code, because that is what Navid runs his own business on and it gives you the most leverage today.

The structure itself is deliberately tool-agnostic. Your context, voice and workflow files are plain files you own, so they work with the others now and carry over when a better model arrives.

No. Your AI OS is a folder of plain markdown files, so any agent that can read a folder can run on it.

Codex reads AGENTS.md. Cursor reads AGENTS.md and CLAUDE.md. Gemini CLI reads GEMINI.md. In the kit all 3 are symlinks to the same CLAUDE.md, so the rules can never drift between them.

The plugin route and the brand-guardian agent are Claude Code features, so that is where you get the most out of it.

Claude Pro is enough for almost everyone starting out.

Claude Code is included with the Pro plan. You do not need Max to set up an AI OS or to run the skills that ship with it.

Max buys you more usage, not more capability. Start on Pro and move up only when you actually hit the ceiling.

A skill is a job your AI OS already knows how to do, invoked by name with a slash command.

The free kit ships 12. A morning brief, a daily wrap, meeting notes, research, slide decks, visual explainers, diagrams, connecting your tools, auditing your setup, keeping the OS updated, and growing it from what you actually did that week.

Each one is a markdown file you can open and edit. They follow the Agent Skills standard, so they are not locked to 1 vendor.

Connectors are the links between your AI OS and the services you already use, so it reads your real data instead of asking you to paste it in.

Think mail, calendar, drive, docs, sheets, notes, your publishing platforms and your analytics. A connect command walks you through each one and records what the OS can reach.

In the free kit you wire these up yourself. In AI OS Pro they arrive already connected.

brand-guardian is a critic agent that checks every draft against your voice before you ever see it.

It reads your voice profile and your list of words you never use, then rejects the drafts that miss. That is what stops the output sounding like generic AI writing.

It ships with the free kit and it is a Claude Code feature.

You can run the work you currently redo from scratch every week.

• Content in your own voice, across every platform you publish on.
• Research and briefs prepared before you sit down.
• Customer replies drafted from your own past answers.
• Slide decks, diagrams and visual explainers from a description.
• Meeting notes filed with the decisions pulled out.
• Daily operations that carry on while you are recording or in meetings.

None of that needs a new prompt. It needs your business written down once.

Start with the thing you rewrite most often.

For most creators that is content in their own voice. For most founders it is customer replies or a weekly brief.

Build that 1 workflow, run it for a week, then add the next. An AI OS grows by use rather than by planning.

Corrections land in files that load at the start of the next session, so a correction you give once does not come back.

What you decide goes in a decisions log. What came up while working goes in memory. What is still open goes in an open work file.

That is the difference between a chat and an operating system. The chat forgets on close. The OS reads its own notes on open.

It reads your real writing during setup, then keeps 2 files that every draft is checked against.

The first is a voice profile: how you actually sound, drawn from your own sent mail and posts rather than from a description of yourself. The second is a list of words you never use.

Both are read before a single sentence is drafted, and the brand-guardian agent throws out the drafts that miss. You can edit either file whenever your voice moves.

Because most people quit for the same reason, and an AI OS removes it.

Every session starts from nothing, so the effort never accumulates and it always feels like starting over. That is exhausting, and quitting is the rational response.

The context you write once is read at the start of every session from then on. The second week is better than the first. The second month is better than the second week.

It grows from what you actually do, not from a roadmap you write in advance.

A grow command looks at your week and proposes the next skill based on the work you repeated. An update command keeps the OS current as the agents change. An audit command tells you what your setup is still missing.

Add 1 thing at a time and run it for a week before you add the next.

AI OS Pro is a complete AI Operating System that arrives already built.

The skills are written, your tools connect properly, and routines run on a schedule. It is the full system Navid runs his own business on, rather than a course or a download.

You install it on your own computer the same way you install the free kit. Everything stays on your machine.

The free kit gives you a system that knows you. AI OS Pro gives you a system that reaches your data and acts on it.

The free kit ships 12 skills and a connector file you fill in yourself. Pro arrives with the skills already written, the connectors already wired, and routines that run whether or not you open the laptop.

Pro is not an add-on to the free kit. It is the same idea, finished.

AI OS Pro includes 4 things.

• The AI OS workspace template and guide. The complete workspace Navid runs his business on, with CLAUDE.md, context files, voice files, starter skills and plugins, plus a written guide and a video walkthrough.
• The Claude Skills Library. Tested Claude Code skills and plugins built for creators and founders, ready to drop straight in.
• The AI Connectors Vault. A growing library of MCPs and CLIs that connect your agent directly to the platforms you run your business on.
• The AI Creator Prompt Library. 100+ proven prompts for content, products, offers and campaigns.

No. AI OS Pro installs on its own.

The free kit at aios.guide is there if you want to try the idea before you buy. If you already built one, the Pro setup reads what you have written instead of overwriting it.

Starting free and upgrading later costs you nothing but time.

AI OS Max is everything in AI OS Pro plus the AI OS Command Center.

Those are the 2 paid rungs. Pro is the system. Max is the system with a cockpit on top of it.

If you are not sure which you want, Pro is the one to start on.

The AI OS Command Center is your mission control for your AI OS: daily plans, live metrics, agent actions and 1-click workflows in a single panel.

It shows what is connected, what ran, what it cost and what is queued. It runs on your machine and reads your folder.

It is included in AI OS Max.

AI OS Pro is a one-time purchase rather than a subscription.

There are 2 rungs. AI OS Pro, and AI OS Max, which adds the AI OS Command Center. The current figure for each is shown at checkout, because it moves with whichever offer is running.

The free starter kit stays free either way.

Yes. The folder is on your computer and the version you have is yours to keep, with no expiry.

Nothing switches off, nothing phones home, and nothing stops working. What stops is access to new versions.

That is the point of an AI OS being plain files. Nobody can take it back.

Yes. Every file is plain markdown on your own computer and all of it is yours to read and change.

Pro is licensed rather than sold, so you use and edit it for your own business. You do not republish or resell it.

The free starter kit is MIT licensed and carries no such limit.

No. Your AI OS is a folder of files you control, on your computer and in your own GitHub repo if you use one, and nothing in it is uploaded to us.

There is no backend, no account and no sync. Nobody here ever sees what you write in it.

Your agent reads those files the same way it reads any other folder you open, so the only thing that leaves your machine is what you choose to send to your own AI provider under your own account.

Your keys live in .claude/settings.local.json inside your own folder, and that file is in .gitignore.

That means a key can never be committed by accident, even if you push your AI OS to a repository. The .gitignore also catches every .env variant, including backups.

Check it yourself before you connect anything. It is 1 file in the repo.

Keep it private, or keep it off git entirely.

Your context folder holds your business: who your customers are, what you sell, and how you write. That is not something to publish by accident.

The install makes the folder a git repo of your own so you get version history. Just do not make that repo public.

You get the starter kit, then the 7-lesson course, then Navid's emails about building and running an AI OS. Nothing else.

Your address is never sold and never shared.

Every email has a 1-click unsubscribe that works immediately.

Yes. The kit is updated as the agents change and as better patterns show up in Navid's own AI OS.

Every change is in the changelog, newest first, and releases are tagged. So you can see exactly what you have and what has changed since.

AI OS Pro is updated on the same basis.

Run the update command inside your AI OS and it pulls in what changed.

It updates the OS itself and leaves your own files alone. Your context, your voice profile, your memory and anything you have written stay exactly as they are.

Read the changelog first if you want to know what is coming before you run it.

On GitHub. The kit is public at thenavidm/ai-os.

Read every file, check the .gitignore, look at the 12 skills, and see the MIT licence for yourself. There is nothing hidden and nothing minified.

If you would rather look up a word first, the AI glossary explains every term used here in plain language.

Yes. Use the support link in the footer, or open an issue on GitHub.

Navid and his team get back to you within 24 to 48 hours.

If your question is about a word rather than the product, the AI glossary is faster.