A self-expanding OS for you and your AI agents. Build apps, connect tools and keep the work on your own private cloud computer.
See how it works
Always-on computers start at $20/month. Add capacity when your work grows.
The idea
Matrix OS is a self-expanding operating system on a private cloud computer for you and your AI agents. It brings compute, integrations and inference into one workspace: build your own apps, connect your tools, chat with supported agents and use a real terminal. Every app you add becomes another part of your computer.
14-day trial · Card required · Cancel before renewal
A place to begin.
01 A computer
A conversation is where a task begins. A computer is where the work happens: a terminal, a project, the tools it needs, and somewhere to save the result.
Matrix brings that environment together on your own private cloud computer.
Meet your computer02 Your agents
Use Claude Code, Codex, or another supported agent with a real terminal and project files. Install the tools your work calls for, rather than rebuilding your environment around every task.
Use separate sessions and Git worktrees when agents edit in parallel. Review the changes before you ship.
Explore supported agents
03 Connected tools
Your projects already span more than one app. Connect supported tools such as GitHub, Gmail, Calendar, Drive and Slack, then give your agent a task with a clear outcome.
Each connection uses the access you grant. Start with the one your first task actually needs.
See the integrations
04 Work you keep
A useful result can be a brief in your files, a change in your repository, or an app you use every day. Keep the project beside the tools and instructions that helped create it.
Save what worked as a project note or skill. Your next task can start with the standards you chose.
See how context stays
05 Keep going
Your Matrix computer runs in the cloud. Leave a supported background task there, close your laptop, and return to its files and results from your next session.
Always-on compute is available from the $20 plan. Running agents still depend on valid credentials, available AI usage and the limits of their tools.
Find your cloud planOne computer.
More room for what you do next.
In practice
Start with a real outcome, connect only what it needs, and decide where the agent should stop for review.
Describe a habit tracker, content planner or client dashboard. Your agent builds it in your workspace. Try it, ask for changes and keep its data with the rest of your work.
A useful first task for anyone getting started.Give Hermes a bounded recurring job: gather the updates from the tools you connected, assemble a report in your workspace, and flag the decisions that still need you. Review the source material before you act.
Scheduled background work requires Builder or Max.Bring a transcript or meeting notes into the computer. Ask your agent to find the strongest ideas, prepare a brief, and draft the next assets in your own files. Save your edits as guidance for the next round.
Your voice and final approval stay with you.Clone a repository, install your tools, start a service and open a named terminal session. They belong to the computer, not to a single chat. Leave one supported client and return to that same environment.
Saved files remain when a process stops. A running agent still depends on the computer, its credentials and its own provider limits. Matrix makes those boundaries visible so you can inspect the actual result.
Claude Code, Codex, OpenCode and other compatible native tools can work in the same computer. Use separate branches or worktrees when they edit in parallel, then review the files and tests yourself.
Matrix Cloud runs the computer for you. The runtime is open source, and manual self-hosting is in Preview. Cross-provider task history, review and recovery are still developing.
Open your Matrix computer in the web app or desktop app. Connect from your own terminal with the CLI. On a phone, use the browser workspace to check in; the native mobile app is in early access.
For messaging workflows, configure Hermes or OpenClaw on your computer and connect a supported channel. Channel availability and model setup depend on the runtime you choose. Your saved files and sessions remain on the Matrix computer.
Your call
Choose which services to connect and what a workflow may do. Ask for drafts before messages or posts go out. Keep consequential actions behind a review step.
Inspect the computer, the source files, and the outputs yourself. Use separate branches for parallel coding work. Agent and provider limits still apply, and every important result deserves a human check.
Built in Europe
Matrix is built in Europe and supports people and teams in the EU and UK. Choose an available compute region during setup, including German locations for the computer itself. Your selected region appears in your computer settings.
Region selection describes where the computer runs; it is not by itself a guarantee about every connected service, model provider, or legal data residency requirement.
Explore regions and setupComing soon. We’re building our own AI gateway for Matrix OS, and the waiting list is open.
Today, use supported agents with your own provider accounts. AI subscriptions and usage are billed separately from your Matrix computer.
Explore the AI gateway waitlistA small, reviewable result is the quickest way to see whether a persistent computer belongs in your workflow.
Open Matrix and choose a computer.
Describe an app you want to use and connect your preferred agent.
Use the first version, review the result and ask for improvements.