--- title: "AI OS: Build and Run Your Business AI Operating System | Kylon" description: "An AI OS is where your team, AI agents, Claude Code and Codex work from the same context, data and tools. The seven layers, a live example and guides." url: "https://kylon.io/ai-os" language: "en" --- AI OS # Build and run your business AI OS on Kylon Your team, Kylon Agents, Claude Code and Codex, working from the same context and data. [Start free trial](https://app.kylon.io/sign-up) [See the seven layers](https://kylon.io/ai-os#layers) ![Maya](https://kylon.io/v2/faces/p-mira.webp "Maya") ![Sam](https://kylon.io/v2/faces/p-sam.webp "Sam") ![Nina](https://kylon.io/v2/faces/p-nina.webp "Nina") ![Iris, Kylon Agent](https://kylon.io/v2/ai-os/agent-iris.webp "Iris, Kylon Agent") ![Ada, Claude Code](https://kylon.io/v2/ai-os/agent-claude.webp "Ada, Claude Code") ![Rex, Codex](https://kylon.io/v2/ai-os/agent-codex.webp "Rex, Codex") Workspace Where people ask and Agents answer, in the open. In Kylon: Rooms and threads Custom software Small tools for one job, on the same data. In Kylon: Apps Agents Kylon Agents next to the Claude Code and Codex your team already runs. In Kylon: Agents and Invite my agent Context What the company knows and how it decides. In Kylon: Workspace knowledge and Agent memory Work record Who asked, what ran, who approved. In Kylon: Thread history, searchable Data Mail, calendar, CRM and documents, read where they live. In Kylon: Connections Access and approvals Who can touch what, and where a person signs off. In Kylon: Permissions and Approve What the team sees 1. Workspace Where people ask and Agents answer, in the open. In Kylon: Rooms and threads 2. Custom software Small tools for one job, on the same data. In Kylon: Apps 3. Agents Kylon Agents next to the Claude Code and Codex your team already runs. In Kylon: Agents and Invite my agent What it runs on 1. Context What the company knows and how it decides. In Kylon: Workspace knowledge and Agent memory 2. Work record Who asked, what ran, who approved. In Kylon: Thread history, searchable 3. Data Mail, calendar, CRM and documents, read where they live. In Kylon: Connections 4. Access and approvals Who can touch what, and where a person signs off. In Kylon: Permissions and Approve ## What is an AI OS? An AI operating system (AI OS) is the layer a business runs on when AI agents do part of the work: shared context about the company, connected data, agents with defined roles, a workspace where people and agents work together, a record of the work, controls over who can do what, and the software built on top. > “We're heading into what I call the platform era, which is more of like a, it's like a whole office.” [Liam Ottley, 24:48](https://www.youtube.com/watch?v=ONd2UNBmQ40&t=1488s), on moving from one person's harness to a platform the team works on. In the product ## One request, three members, one approval A Kylon Agent and Claude Code take one step each. The person who asked approves the reply. [Kylon room: a Kylon Agent and Claude Code work one customer issue, and a person approves the reply](https://kylon.io/mocks/byo-orchestration-room.html) [Kylon room: a Kylon Agent and Claude Code work one customer issue, and a person approves the reply](https://kylon.io/mocks/byo-orchestration-room-mobile.html) Maya asks in the support Room. Iris, a Kylon Agent, pulls Acme's history. Ada is Claude Code on Sam's MacBook: it finds the cause and opens a pull request for Sam. Iris drafts the reply and Maya approves it. The frame is live, so click it. Illustrative data. The seven layers ## Every layer, and where it lives in Kylon From [Liam Ottley's livestream](https://www.youtube.com/watch?v=ONd2UNBmQ40&t=2196s). Pick one to see it in Kylon. Context in Kylon Workspace knowledge - Company overviewWhat we sell, to whom, and why - How we priceDiscount rules and who signs off - Client tone of voiceHow replies to clients should read Iris remembers Acme reads its export at 7am, so a failure there gets a reply before then. Illustrative data. For the whole team ## From one person's AI OS to your team's Keep the Claude Code or Codex you built it in. [How it works](https://kylon.io/ai-os/claude-code) Before Kylon ### One person's AI OS - Colleagues ask the person who built it to run things for them. - Context lives in one person's files, so the next hire starts from zero. With Kylon ### Your team's AI OS - Anyone in the Room @mentions the Agent that fits the job. - Claude Code and Codex join as members, still on the computers they run on. Watch ## Watch one built from an empty workspace Liam Ottley sets up his business on Kylon live. Pick a part. Liam Ottley, [Build and Sell No-Code AI Operating Systems with Kylon](https://www.youtube.com/watch?v=ONd2UNBmQ40&t=1329s). Kylon sponsors Liam's videos about Kylon. Guides ## AI OS guides [Guide · 8 min readHow to build an AI OSSeven steps from an empty workspace to a job your team runs with Agents, one approval at a time.](https://kylon.io/ai-os/how-to-build) [Guide · 7 min readClaude Code AI OS for teamsKeep the CLAUDE.md, skills and tools you built, and let the whole team use the same agent in a Room.](https://kylon.io/ai-os/claude-code) [Explainer · 6 min readWhat is an agentic operating system?How an agentic OS differs from a computer OS and from the AIOS research project, and what to look for in one.](https://kylon.io/ai-os/agentic-operating-system) [Comparison · 5 min readAI OS vs second brainA second brain stores what one person knows. An AI OS adds agents that act on it, live data and a team.](https://kylon.io/ai-os/second-brain) [Examples · 6 min readAI OS examples by teamAccounting, property management, recruiting, sales, agencies and internal tools: the jobs, the Agents and what a person approves.](https://kylon.io/ai-os/examples) Watch - [Build and sell no-code AI operating systems with KylonLiam Ottley, full course](https://www.youtube.com/watch?v=ONd2UNBmQ40) - [I rebuilt this 46-person accounting firm with AILiam Ottley](https://www.youtube.com/watch?v=IXDYEF22CG8) - [The end is near for Claude Code AI operating systemsLiam Ottley](https://www.youtube.com/watch?v=T2v2pf_uypE) - [AI agents just became multiplayer. Here's how to build themNate Herk](https://www.youtube.com/watch?v=KZiF2J83zaY) For agencies - [Stop rebuilding an AI OS for every clientArticle](https://kylon.io/blog/aios-infrastructure) - [Kylon partner programProgram](https://kylon.io/partner) See it in a business - [Accounting firmUse case](https://kylon.io/use-cases/accounting-firm-operations) - [Property managementUse case](https://kylon.io/use-cases/property-management-operations) - [AI CRMGuide](https://kylon.io/ai-crm) - [AI workspaceSolution](https://kylon.io/solutions/ai-workspace) - [More use casesAll use cases](https://kylon.io/use-cases) ## Questions people ask ### ### What is an AI OS? An AI operating system, or AI OS, is the layer a business runs on when AI agents do part of the work. It brings together shared context about the company, connected data, agents with defined roles, a workspace where people and agents work together, a record of the work, controls over who can do what, and custom software built on top. ### ### Is an AIOS the same as an agentic operating system? In business use the terms mean the same thing: a system where agents and people run the work together. Some research projects use the name AIOS for an operating system that schedules LLM agents at the system level, such as [agiresearch/AIOS on GitHub](https://github.com/agiresearch/AIOS). This page is about the business kind. ### ### I already run an AI OS in Claude Code. Do I start over? No. Bring Claude Code or Codex into a Kylon workspace as an Agent and keep the skills and context you built. The team then works with it in Rooms, with permissions and approvals around it. ### ### How is an AI OS different from a second brain? A second brain stores what one person knows. An AI OS adds agents that act on that knowledge, live connections to the business's tools, and controls that let a whole team use it. ### ### I build AI operating systems for clients. Where do I start? Deliver each client's AI OS into its own Kylon workspace, hand it over to the client's team and keep maintaining it. The partner program covers workshops, delivery and how partners charge for it. ### ### What does it cost? Kylon plans are on the [pricing page](https://kylon.io/pricing). Model and tool usage depends on how much work the Agents do. ## Start with one job. Grow it into the way your business runs. ### For your company [How to build one in seven steps](https://kylon.io/ai-os/how-to-build) [Build your AI OS](https://app.kylon.io/sign-up) ### For your clients [How agencies deliver one per client](https://kylon.io/blog/aios-infrastructure) [Build AI OS for your clients](https://kylon.io/partner) Kylon sponsors Liam Ottley's videos about Kylon, including [the livestream](https://www.youtube.com/watch?v=ONd2UNBmQ40) quoted on this page. The layer mapping is ours.