OpenMausBot: an open-source team of AI bots in a chat app

A roster of real AI agents in a messaging app - each with its own model, computer and apps - running locally on Claude, Codex or Grok. The open-source alternative to Grok Bot, Muse, dots and Cue.

12 min read

OpenMausBot is an open-source Grok Bot alternative: a chat application where every conversation is a real AI agent. Each bot in the sidebar is Claude, Codex or Grok running locally under the hood, with its own personality, its own model, its own cloud computer and its own connected apps. You talk to them like contacts, watch them work, and approve what matters. It is local-first, licensed under Apache 2.0, and runs on the agent CLIs you already have installed.

What OpenMausBot is

One assistant in one box is the wrong shape for agents. OpenMausBot keeps the idea of AI as a messaging app - a roster of bots you chat with, each with its own personality, memory of its thread, model, computer and apps - and rebuilds it open, local-first, on the agents you already have.

  • Bring your own agents: bots run on the claude, codex and grok CLIs installed on your own machine - your existing logins and subscriptions, no new accounts, no proxy in the middle. Point any engine at a custom CLI binary in Settings - Engines.
  • Local first: one small harness server on 127.0.0.1 owns every agent process. Transcripts, keys and events live in ~/.openmausbot, not a cloud.
  • Agents with hands: each bot can use a cloud Linux desktop, an isolated local VM, or - where the platform safety boundary is certified - your own computer, plus 500+ apps through Composio.
  • Approval cards: shell commands, file edits and questions surface as inline Allow / Deny / answer cards, for cloud and local computers alike.

The open-source alternative to Grok Bot, Muse, dots and Cue

Four closed personal-agent products shipped in seven weeks of 2026, and each one keeps the model, the computer and your data on its maker's side of the line. OpenMausBot is the open-source version of that shape: a team of agents in a chat app, each with its own model, its own computer and its own connected apps, running on your machine under Apache 2.0.

  • Open source Grok Bot alternative: xAI's roster of bots on one shared cloud computer, Grok only, from a SuperGrok or Cursor plan. OpenMausBot keeps the roster shape but allows any model per bot, runs on your own machine, and uses approval cards.
  • Open Muse alternative: Meta's single personal agent in a Meta cloud VM, US and Canada only. OpenMausBot is a team instead of one agent, runs anywhere, and keeps data in ~/.openmausbot.
  • OpenAI dots alternative: OpenAI's always-on agent inside ChatGPT, GPT only, Pro and Business Premium plans. OpenMausBot runs several agents on your Claude, ChatGPT or Grok login with no plan required.
  • Open Cue alternative: Manus's invite-only agents with their own phone, email and wallet. OpenMausBot needs no invite, signs in over OAuth, and is open source.
  • Open Instinct alternative: Spear Street's invite-only agent that holds your passwords on a cloud computer. OpenMausBot asks for approval when you want it and signs in over OAuth on your machine.

All five side by side are compared on the OpenMausBot site, along with a decision guide in five questions: which AI agent should I use?

Key features

Pick a brain per bot

A model picker with a provider rail shows Claude and Codex models side by side, defaults marked, unavailable providers dimmed with the reason. You can switch a bot's model mid-conversation.

Every bot gets a computer

Open the Computer panel and the bot's cloud desktop spins up on its own - live screen preview while it works, Open desktop to take over in your browser, or point the bot at this Mac instead.

Bots ask before they act

Shell commands, file edits and questions surface as inline cards - Allow, Deny, or answer in chat. A permission broker turns every risky action into a decision you make, for cloud and local computers alike.

Connected apps

A one-click marketplace over Composio Sessions: Gmail, Slack, GitHub, Notion, Linear and hundreds more. OAuth once, and every bot can use them as tools.

Manage bots like chats

Right-click any bot to pin, mark unread, edit profile, duplicate, copy conversation ID, hide or delete. It is a messaging app - your agents behave like contacts.

Channels for every context

Keep Work, Personal and each project in separate channels without cloning your bots. Every channel has its own transcript, shared instructions, working folder, responder rules and editable bot roster.

Fast routing with Jev

Add a Cloudflare Workers AI token with its account ID, a TypeSafe Jev key, or an OpenRouter key and model, and rooms set to Auto pick the bot that fits each un-mentioned message in a few hundred milliseconds, falling back to the room's lead whenever the decision model is off or unsure.

Install a complete team from one Markdown file

Browse outcome-driven teams on BotMRR, then choose Add to OpenMausBot. The app opens a review screen before creating the bots, Chief of Staff, channels, playbooks, connector checklist and suggested routines. You can also import the same .md file from disk or paste its public GitHub URL in Teams - Import. Connections stay off until you approve them, routines arrive paused, and packages never carry credentials, conversations, permissions, memory or computer access.

Share a whole department

Right-click a department and choose Share department to save the whole department as one file: its bots with instructions, pictures and skills, group chats, routines and shared instructions. Chat history, keys, passwords, model choices and computers never go in, and anything that looks like a secret is removed from the text. Whoever imports it gets new bots on Ask, skills switched off and routines paused.

Bots that talk back

Press the speaker on any reply, or switch a bot to read its answers out as they land. Hit call and it is a conversation: it hears you, tells you what it is doing while it works, and asks for approvals out loud. Voices come from ElevenLabs, Fish Audio, Grok (xAI), built-in Mac voices or a local Chatterbox server.

Triggers: schedules and webhooks

A trigger is when something happens, a bot does something. Tell a bot every weekday at 9, send me a summary of my inbox, or whenever our form calls, draft a reply, and it sets the trigger up in that turn with its trigger tool. Schedules can run once, on selected weekdays, or every 5 to 1440 minutes. Webhook triggers are independent but reuse the same queued task executor. App events start a bot when something happens in a connected app, such as an email from Acme.

How it works

Two processes. The app holds no transports of its own - it sends typed commands over HTTP and folds one SSE event stream into state. The harness server owns every agent process and normalizes each provider's native protocol into one canonical runtime event stream, logged per thread as NDJSON.

  • Drivers in server/drivers/: one per provider - Claude, Codex and Grok Build over their local CLIs (stream-JSON, JSON-RPC, ACP), plus a cloud-computer agent. Unknown drivers degrade to unavailable, never crash the fleet.
  • Harness in server/harness/: registry (configs to live instances) and the fan-in event bus every client folds.
  • API in server/index.ts: bots, turns, approvals, model catalog, computer lifecycle, connectors, config - HTTP plus SSE.
  • Voice in server/tts/: ElevenLabs, Fish Audio, Grok (xAI), built-in Mac voices or local Chatterbox. Cloud keys stay on the harness.
  • App in src/: the chat shell. Server-backed store, one reducer, zero client-side transports.
  • Desktop in electron/: macOS, Windows and Ubuntu shells with an embedded harness and platform capabilities.

OpenMausBot also ships a stdio MCP server for external clients such as Claude Desktop and Cursor. It exposes a deliberately bounded team control plane: inspect bots and channels, read and search compact transcript pages, create and configure bots, channels and tasks, send work, wait for completion, switch models and interrupt turns. It does not expose approval grants, deletion, arbitrary settings, credentials or computer lifecycle.

The openmausbot command drives a running workspace without a screen: set a team up from a file, give a bot one job, wait until it and every teammate it asked are done, and export what they made. One JSON document with --json, and exit codes a script can branch on.

Getting started

  1. Download a released build: macOS (Apple silicon or Intel), Windows x64 or Ubuntu 24.04 x64. The harness server is embedded, so no separate server setup is required.
  2. Or run from source: git clone https://github.com/milind-soni/OpenMausBot, then pnpm install, pnpm dev:server (harness on 127.0.0.1:8799) and pnpm dev (app on 127.0.0.1:5199).
  3. Requirements: macOS, Windows or Ubuntu 24.04 x64, Node 24+, pnpm, and at least one agent CLI - claude, codex or grok - installed and logged in. They appear in the model picker automatically.
  4. First launch guides you with arrow-key choices: choose AI access, sign in or paste a hidden API key, choose a model, and optionally connect a phone.
  5. Optional credentials: Composio key for connected apps, Boat API key for cloud computers, ElevenLabs or Fish Audio keys for voices.

Desktop capability status

  • Packaged app, embedded harness, local agent CLIs: supported on macOS and Windows, beta on Ubuntu.
  • Composio and Boat cloud computers: supported on macOS and Windows, beta on Ubuntu.
  • Bot control of this computer: supported on macOS and Windows, beta with explicit opt-in on Ubuntu Xorg, disabled on Wayland.
  • Native on-device dictation: supported on macOS, unavailable on Windows, planned on Ubuntu.

Status and rough edges

Early but real - the loop works end to end: message to agent to streamed reply to tools to approvals to computer use. macOS, Windows and Ubuntu 24.04 x64 have released builds; Ubuntu remains a beta. Rough edges to expect: hosted and mobile connectivity is still being built, and webhook triggers currently use the local receiver rather than an always-on hosted relay. Hosted voice needs an ElevenLabs, Fish Audio or xAI key; built-in Mac and local Chatterbox voices need no cloud key.

Contributions are welcome. The driver SPI in server/contracts.ts is deliberately small; adding a provider is one file in server/drivers/ plus a one-line registration. No code is needed for your own engines: any ACP-speaking CLI or OpenAI-compatible endpoint plugs in through config.

Frequently asked questions

Is OpenMausBot really free?

Yes. Free and open source under Apache 2.0. Cloud computers via Boat and hosted voices are paid services with their own accounts, but local chat works without them.

Does it send my data to a cloud?

No. Transcripts, keys and events live in ~/.openmausbot on your machine. Usage analytics are optional and anonymous, and can be turned off in Settings.

Which agents does it support?

Claude, Codex and Grok, through their local CLIs. Any ACP-speaking CLI or OpenAI-compatible endpoint can also plug in through config, and users can add their own MCP tool servers with zero code.

Can an AI agent drive OpenMausBot itself?

Yes. OpenMausBot ships a stdio MCP server that lets external clients inspect bots and channels, send work, wait for completion, switch models and interrupt turns.

Where is the source code?

On GitHub at github.com/milind-soni/OpenMausBot, with a separate releases repository for signed macOS builds.

One assistant in one box is the wrong shape for agents.

— OpenMausBot

If you found this useful, browse the rest of the YesIntelligent articles for more practical guides on creator tools, open-source software and reading engagement honestly.

Related articles

Back to all articles · Browse all tools