OpenMausBot Runs a Team of AI Agents in One Desktop App

Running several AI agents usually means a stack of terminal windows and no clear picture of what any of them is doing. OpenMausBot is an open-source, local-first desktop app that gives each bot its own chat thread instead.

Every agent gets its own conversation, model, computer, and connected apps, so you can hand out work and follow it in one place. It runs on your machine, not someone else's server.

What Happened

OpenMausBot arrived as an open-source desktop app for running a team of AI bots in a chat-style interface. It is local-first, meaning the agents run through CLIs already installed on your computer, and you bring your own: it works with Claude, Codex, or Grok.

Each bot gets its own personality, model, computer, and connected apps, so two agents can be set up for two completely different jobs.

For the computer side, a bot can use a cloud desktop, an isolated local VM, or supported host control. Shell commands, file edits, and questions from a bot surface in the chat as Allow / Deny cards before anything happens.

Context channels keep work, personal, and project conversations apart, each with its own instructions and its own bot roster.

The open-source edition is licensed under Apache License 2.0. The enterprise/ directory has its own separate license.

Why It Matters

If you already run agents from the command line, this is the part that changes your day: instead of guessing which window is doing what, each agent reports in its own thread. The inline approvals are the other piece worth noticing. A shell command or a file edit stops and waits for your yes, so a bot cannot quietly run something you never agreed to.

The tradeoff is that it depends on CLIs you install yourself, and connecting outside apps is part of setup rather than something that happens on its own. For anyone coordinating more than one agent at a time, that is still a shorter path than keeping a mental map of open terminals.