NAISYS

NPM | Website | Discord | Demo Video

Self-hosted, cross-machine, AI agent runner and manager that runs on Node.js

• Collaborative — Invite others with fine-grained permissions to manage agents, teams, and machines together
• Multi-agent — Organize agents into teams, communicate via mail, wake on messages, and scale across machines through a self-hosted hub
• Computer use support — In addition to the shell, agents can control any Windows/Mac/Linux GUI
• Cost controls — Costs tightly tracked, requiring a spend limit (fixed or rolling window), no uncapped runs
• All vendors supported — OpenAI, Google, Anthropic, OpenRouter and local LLMs
• Remote management — Manage agents, hosts, costs, and messaging through your self-hosted server

Packages

NAISYS is split into four packages: naisys (the agent runner), @naisys/hub (central server for persistence and cross-machine communication), @naisys/supervisor (web UI for monitoring and managing agents), and @naisys/erp (AI optmized task management system).

Getting Started

Ephemeral Mode

Lightweight agent runner with no persistence or web UI. Runs agents on demand. Pass a directory to run all agent yamls in that folder. Make sure you're running at least Node.js 22 node -v

npm install naisys
npx naisys agent.yaml

The setup wizard on first run will guide you through creating a .env and agent.yaml file.

Integrated Mode

Everything in a single process — Hub for persistence, Supervisor web UI, and optional ERP. Requires a local install since multiple packages are needed:

npm install naisys @naisys/hub @naisys/supervisor @naisys/erp
npx naisys --integrated-hub --supervisor --erp

When the process starts up it gives you a link to the supervisor UI and a command to change the superadmin password.

Simple Client/Server with ngrok

Host a NAISYS server on your own machine and connect remote machines to it through ngrok. This gives you remote management for free with no central server required. Best practice is to run NAISYS clients and servers from within a VM on their own NAISYS account.

Start with the integrated server command above, and then expose the server with ngrok:

ngrok http <port of your naisys instance>

Then on each client machine run this command. This will allow NAISYS to start/stop agents on the machine.

npm install naisys
npx naisys --hub=https://<server>/hub

That's it — you now have a self-hosted cluster of machines to run agents across. The agents can securely communicate with each other, all managed through the supervisor AI.

Advanced Client/Server with PM2

PM2 keeps NAISYS running through system restarts, NAISYS upgrades, and unexpected crashes. NAISYS server uses around 180MB, so if you want you could easily host it on a $6/month Digital Ocean Droplet VM.

Create a dedicated user on server/clients for NAISYS

npm install -g pm2
sudo useradd -m -s /bin/bash naisys
sudo su - naisys

Install NAISYS on the server like above then run:

pm2 start npx --name naisys-server -- naisys --integrated-hub --supervisor --erp

Similarly on each client install NAISYS then run

pm2 start npx --name naisys-client -- naisys --hub=https://<server>/hub

Finish PM2 setup to ensure it starts NAISYS on boot

pm2 startup
pm2 save

Installing/Running from Source

git clone https://github.com/swax/NAISYS.git
npm install && npm run build
mkdir testrun && cd testrun
node ../apps/naisys/dist/naisys.js --integrated-hub --supervisor --erp

NAISYS ERP

A light free ERP system optimized for agents by utilizing advanced HATEOAS for API self discovery as well as optimized request/response to minimize payloads. NAISYS ERP enables the controlled release of instructions for agents to follow, helping to minimize variation in task results. A completely optional package that can be run in-process or standalone.

Additional Information

• See agents/template.yaml for all agent configuration options and supported models
• See apps/naisys/README.md for detailed NAISYS documentation
• See apps/supervisor/README.md for Supervisor documentation

Requirements

• Node.js 22 or higher
• Linux, Mac, Windows
• lynx browser: apt install lynx (Linux) or brew install lynx (macOS)
  • For text web browsing, in desktop mode NAISYS can use graphical web browsers

License

MIT