OpenMausBot: A local-first chat app for running a team of AI agents.
What problem it solves
Teams often need several agents with distinct personalities, models, computers, and connected apps rather than one assistant in a single interface. These agents must also work with existing local CLI sessions, keep local data under user control, and request approval before risky actions.
Key features
Bring your own agent engines — Run bots through Claude, Codex, or Grok CLIs installed on the local machine, using existing logins and subscriptions. An engine can also point to a custom CLI binary in Settings → Engines.
Source: README
Give every bot a computer — Each bot can work in a cloud Linux desktop, an isolated Local VM, or an eligible local computer. The Computer panel includes a live preview and browser access for taking over the desktop.
Source: README
Approve risky actions — Shell commands, file edits, and agent questions appear as inline cards that can be allowed, denied, or answered in chat. The permission broker applies these decisions to cloud and local computers.
Source: README
Connect apps and organize contexts — Composio Sessions provide access to Gmail, Slack, GitHub, Notion, Linear, and other apps through OAuth. Work, personal, and project contexts can each maintain a transcript, instructions, working folder, responder rules, and editable bot roster.
Source: README
Automate from scripts and CI — The headless command can apply a team configuration, assign a job to a bot, wait for the bot and its teammates to finish, and export the results as JSON or other output.
Source: README, docs
Quick start
Copied verbatim from the project's README.
git clone https://github.com/milind-soni/OpenMausBot && cd OpenMausBot
pnpm install
pnpm dev:server # harness server → 127.0.0.1:8799
pnpm dev # app → http://127.0.0.1:5199
pnpm dev:desktop # Electron shell; keep the two commands above runningopenmausbot team apply team.yaml
openmausbot run "Room Lead" "Rehearse the launch of https://example.com" --export ./outWho it may suit
This project may suit developers and teams who want to organize multiple local AI agents as chat contacts and retain explicit control over tools, approvals, and connected apps.
Topics to discuss
How should local agent permissions be designed across cloud, VM, and host-computer environments?
Which workflows would be most useful to package as portable Markdown team templates?
How can MCP integrations support agent teams without exposing sensitive control operations?
Written by ReadmeX AI from the project's README, source and public GitHub data; it may contain mistakes.
Read the original README https://github.com/milind-soni/OpenMausBot#readme
ReadmeX opened this community for a trending project; it is not run by the project. Maintainers can claim it from the community page.
Comments
…