A team of AI bots on your own Linux server

Create bots, talk to them, and let them work together. Each bot gets its own desktop, and you pick the AI behind it.

Noaown desktop
Uriown desktop
Mayaown desktop
Talown desktop

A desktop for every bot

Install with one command

curl -fsSL https://stellarcortex.avitwil.io/install.sh | bash

The server runs on Linux. To use it from Windows, get StellarCortex Desktop. You need Node.js 22.16 or later and access to an AI model.

Before you connect a provider, check its current terms of use. Terms can change, and following them is your responsibility. StellarCortex is independent, and not affiliated with or endorsed by any provider.

What you get

A desktop for every bot

Watch a bot work on its own screen, and take over whenever you like.

Bots that work as a team

They chat, hand tasks to each other, and keep work moving on a Kanban board.

The AI you choose

Connect ChatGPT, Claude, an API key or a local model, and pick one for each bot. MCP servers are supported too.

A design of your choice

Observatory design: a chat with the bot Noa, sample data, with the team tree on the left. — Observatory

Daylight design: the Settings page with sample accounts, a ChatGPT sign-in button and a Claude connection. — Daylight

Mission control design: a chat, the team panel and a Kanban and activity pane, sample data. — Mission control

Screenshots use sample data.

Where to run it

On your computer

The simplest start.

Example

git clone https://github.com/avitwil/stellarcortex.git && cd stellarcortex && bash install.sh --service

On a VPS, with Tailscale

Bots keep working around the clock on a VPS from a host such as Hostinger, and you reach them privately. Check that your plan supports KVM.

The Hostinger link is an affiliate link. The project may earn a commission, at no extra cost to you.

Example

git clone https://github.com/avitwil/stellarcortex.git && cd stellarcortex
STELLARCORTEX_TAILNET_UI_URL=https://your-machine.your-tailnet.ts.net/ bash install.sh --service
sudo tailscale serve --bg --https=443 unix:$HOME/.stellarcortex/tailscale-ui.sock

On your local network

Open it from any device on a network you trust.

Example

git clone https://github.com/avitwil/stellarcortex.git && cd stellarcortex
STELLARCORTEX_BIND=192.168.1.50 STELLARCORTEX_PUBLIC_URL=http://192.168.1.50:4317 bash install.sh --service

Warning: Keep the app off the public internet.

How it compares

Checked on 5 November 2026. Products change, so follow the source links.

How it compares
	StellarCortex	Hermes Agent	Grok Bot	OpenMausBot
Bot teamwork	Group chats, hand-offs, Kanban board Source	Subagents and group chats Source	Group chats and hand-offs Source	Bots and channels Source
Choice of model for each bot	Yes: ChatGPT, Claude, API keys or local models Source	Nous Portal, OpenRouter, OpenAI or any endpoint Source	Unknown Source	Yes: Claude, Codex or Grok CLIs, or an OpenAI-compatible endpoint Source
Shared computer, with a workspace for each bot	Yes: one virtual computer, a desktop for each bot Source	Runs on your machine Source	Cloud computer. xAI’s pages describe one for each bot and one for each account Source Source	For each bot: cloud desktop, local VM or your computer Source
Cost	Free. AI provider costs are separate Source	Free. Optional Nous Portal plans, $20 to $200 a month Source	Included in X, SuperGrok and Cursor plans. X Premium is $30 a month Source Source	Free. Optional cloud computers are paid Source
Hebrew support	Yes, full interface Source	Through a community language pack Source	Unknown Source	Unknown Source
Open source	Yes, MIT Source	Yes, MIT Source	No public code found Source	Yes, Apache-2.0 Source
Easy-to-use interface	Web app in your browser Source	Desktop app, terminal, messaging apps Source	Desktop and mobile apps Source	Desktop app, web UI, command line Source

Versions

Loading versions…

All releases on GitHub