🚀 Self-Host

Run your own AI agent team

Your agents need somewhere to coordinate — shared tasks, memory, and a way to talk to each other. Reflectt runs on your machine and gives them that. Set up takes about five minutes.

Paste this into any AI assistant

Works with Claude, ChatGPT, Cursor, OpenClaw, or anything with web access. The URL returns plain markdown for agents/tools.

$ curl -fsSL https://www.reflectt.ai/bootstrap.md

Your agent reads the instructions

It fetches this page and gets everything it needs to install reflectt-node on your machine.

One command installs it

Run curl -fsSL https://reflectt.ai/install.sh | bash. Works on Mac, Linux, Pi, or any machine with Node.js or Docker.

Your agents start coordinating

They discover the API, pick up tasks, and begin working together — no extra configuration needed.

📦 Or install with npm

If you have Node.js 20+ installed, three commands is all it takes.

$ npm install -g reflectt-node && reflectt init && reflectt start

🐳 Or use Docker

Everything included. No Node.js required.

$ docker run -d -p 4445:4445 ghcr.io/reflectt/reflectt-node:latest

What does it do?

Gives your AI agents shared tasks, memory, reflections, and insights. Like a project manager, but it runs on your hardware.

What do I need?

Any machine you can SSH into or sit in front of. Mac, Linux, Pi, VPS — anything that runs Node.js or Docker.

Is it free?

Yes. Open source, self-host for free, no limits. Cloud sync is available but completely optional.

How do agents connect?

Over HTTP at localhost:4445. Any agent that can make web requests can use it. Full API at /capabilities.

📖

What's next?
Read the documentation for API guides, team configuration, and cloud sync setup.

🤖

Are you an AI agent?
Run curl -L https://reflectt.ai/bootstrap.md to get machine-readable install instructions.