OpenClaw: an AI assistant that runs on your own devices

Most AI assistants run on someone else's computer. This one runs on yours, and meets you in the chat apps you already have open.

OpenClaw is a free, MIT-licensed AI assistant that runs on your own devices rather than a vendor's cloud. It connects models, tools and messaging channels through a single component called the Gateway, so the same assistant can answer you in Telegram, Discord, WhatsApp, Signal, iMessage or Slack. The same gateway runs as a personal assistant on one laptop or as a shared deployment for a team that trusts each other, and configuration is the only difference between the two. You install it with a single curl command on macOS, Linux or WSL2, or a PowerShell one-liner on Windows.

What you get

  • An assistant that runs on hardware you control, not a vendor's servers
  • One Gateway connecting your models, your tools and your messaging channels
  • Reaches you in Telegram, Discord, WhatsApp, Signal, iMessage and Slack
  • The same install works solo on a laptop or shared across a trusted team
  • MIT licensed, so you can adapt it and use it commercially

Who this is not for

You are running a service, not opening an app. You need to be comfortable with a terminal, keeping a machine online, and holding your own API keys. If you want something that just works with no maintenance, a hosted assistant is the honest answer.

Credit

Created by OpenClaw Foundation· MIT licensed, free on GitHub

ProfitizeOS did not build this and does not host it. You get a link straight to the original.

Free. Two fields, then straight to the resource and setup guide.

Common questions

Is OpenClaw really MIT licensed?

Yes. GitHub's API reports the licence as NOASSERTION, but the LICENSE file in the repository is a standard MIT licence from the OpenClaw Foundation. The file is what governs.

Does it work on Windows?

Yes, natively. There is a PowerShell installer, and the Linux path also works under WSL2 if you prefer that.

Do I still need API keys?

Yes. Running the assistant on your own device controls where it runs and where your data sits, but the models it calls are still whichever providers you configure, unless you point it at a local model.

Can a team share one instance?

Yes, and this is an explicit design goal, but only for a team whose members trust each other. The gateway does not partition members from one another the way a multi-tenant SaaS would.

More on this topic: AI Automation