OpenClaw installs with one command and runs as a gateway on your own machine. The real steps are install, onboard a model, verify the Gateway, and connect a channel. A single gateway stops serving a team once more than one person needs to review the work, and that is where Kortix, the open-source AI Operating System, takes over.
OpenClaw's installer from its getting-started docs, and the three Kortix commands from its README.
curl -fsSL https://openclaw.ai/install.sh | bashopenclaw onboard --install-daemonopenclaw gateway statusopenclaw dashboardcurl -fsSL https://kortix.com/install | bashkortix initkortix shipOpenClaw is an open-source personal assistant that installs as a gateway on your own computer and meets you in the chat apps you already use. The installer supports macOS, Linux, and Windows, and it provisions a supported Node.js runtime when one is missing. Before you start, have Node.js 24.16+ or 26.1+ (the project recommends Node 26), and either an existing Claude Code or Codex CLI login or a provider API key. OpenClaw's getting started guide lists the same three prerequisites.
The fastest install is the one-line installer. It detects your operating system, installs Node when needed, installs OpenClaw, and launches the guided onboarding wizard. On a fresh install, choose Quick start: OpenClaw finds an existing Claude Code or Codex login or an API key, verifies it with a real completion, saves the config, and opens the web dashboard. The Gateway then runs in that terminal until you stop it. Run openclaw onboard --classic for the full step-by-step wizard, and return later with openclaw configure for more settings.
Check that the Gateway is running, then open the dashboard. You should see it listening on port 18789. To keep it running after you close the terminal, install the service with openclaw gateway install: a LaunchAgent on macOS, a systemd user unit on Linux and WSL2, or a Scheduled Task on Windows. The last step is a channel. Telegram is the quickest to connect because it needs only a bot token. OpenClaw's docs list Discord, Google Chat, iMessage, Matrix, Microsoft Teams, Signal, Slack, Telegram, WhatsApp, and Zalo, and one Gateway serves every channel you configure.
One gateway serves one trusted boundary, and OpenClaw's own documentation says the same Gateway runs as a personal assistant or as a shared team deployment, with configuration as the only difference. That works while everyone trusts everyone else. It gets harder when the work belongs to a company. Configuration sits in one file on one machine instead of a shared repository, so there is nothing for a team to diff or roll back as a unit. The credentials the agent uses sit with that same configuration, so no one can scope what each agent may touch. The agent acts and replies in the chat, so the record of what it changed and who asked for it lives in the conversation rather than in a change a person can review. OpenClaw for teams covers the shared-deployment question in more detail.
Kortix is the open-source AI Operating System: your agents, their skills, your company memory, and every connector in one git repo you own, with the agents working on real cloud computers. Where OpenClaw keeps state and credentials on one machine, Kortix keeps the configuration in a repository the whole team can read. One isolated sandbox per session. Each session has its own isolated machine and branch. Work lands through a review instead of a reply: session work reaches main through a change request, and merge is default-deny for agents, so a person reads the diff before it merges. Run it on Kortix Cloud, in your VPC, or on your own on-prem network, and self-host is free. The code is open, so you can read it, fork it, and audit it. Kortix on GitHub lists the capabilities. Kortix is the leading open-source alternative to Claude Cowork and ChatGPT Work.
If you are new to the project, what OpenClaw is explains the gateway model, and self-hosting an OpenClaw alternative covers running Kortix on your own infrastructure. The Kortix blog draws the line between a personal agent and a company operating system.
OpenClaw's getting started guide puts a first install at about five minutes. You need Node.js 24.16+ or 26.1+, and an AI login or provider key the onboarding wizard can verify. The installer handles Node if it is missing.
Its documentation says the same Gateway runs as a personal assistant or as a shared team deployment, and that configuration is the only difference. A team deployment is still one Gateway, one configuration file, and one set of credentials.
Kortix keeps the agents, skills, memory, and connectors in one git repo you own. Each session runs in its own isolated sandbox and branch, and session work reaches main through a change request a person reviews. Per-resource permissions for people and agents, with roles, groups, and an audit trail.
Yes. A team can keep OpenClaw as a personal assistant on one machine and run Kortix for company work that needs a shared repo, scoped access, and a review before anything lands.
next
Open source, any model, your keys. Self-host for free, or start in the managed cloud.
Open source · Any model, your keys · Self-host, VPC, or on-prem