Local Engine Ready
Early Access Sign Up
Home/Early Access
Early Access // Mac · Windows · Linux Next

Sign up for early access to Celestos.

Celestos is rolling out through private installer waves. Join the list and we will send the right build, setup notes, and release updates as access opens.

Private Builds

One list. The right installer when your wave opens.

We are not opening public downloads yet. Sign up once and we will send the correct Celestos build for your machine, plus setup notes for local models, GPU acceleration, and mobile access.

macOS private installer waves are opening first.
Windows builds are included in the early rollout.
Linux is next and will go to the same signup list.
Sign Up for Early Access
Private Rollout

Get invited when your build is ready.

Tell us where to send access. We will use this for installer invites, launch notes, and practical setup guidance only.

Installer invite matched to your operating system
Setup notes for local models, GPU options, and mobile tunnel
Build updates when new Celestos systems become available

No spam. Just access and meaningful build updates.

Requirements // Before You Install

Runs comfortably on a normal laptop.

A modern CPU with 8GB or more of RAM is recommended — a base model Mac Mini runs it easily
An NVIDIA card with CUDA support speeds things up significantly, but isn't required
The local model runs via llama.cpp — private, free, no API limits
Cloud models are optional and configurable from Settings > Models whenever a task needs more power
What you get on install
Systems 12 specialist systems, one interface
Guides 23 comprehensive guides bundled in-app
Workflows 70+ built-in workflow packs
Mobile Built-in secure tunnel — access from your phone
Mobile // Your PC is the Server

Scan a QR code. Your AI is on your phone.

Celestos includes a built-in secure tunnel that wakes up the moment you start the app. A QR code prints straight in the terminal — scan it with your phone camera and you are in. No account, no pairing, no separate app to install.

A mobile-optimised Orb interface appears at a shareable HTTPS link
Microphone works — voice input is fully supported over the secure tunnel
Your computer stays the private backend — nothing routes through a third-party server
Tunnel launches automatically on startup if the tunneling tool is installed
How it works
Step 1 — Start Celestos on your Mac or PC as usual
Step 2 — A QR code appears in the terminal window automatically
Step 3 — Scan with your phone. The mobile Orb loads instantly in your browser
Step 4 — Speak or type. Responses stream back from your local engine
MOBILE → https://••••.celestos-tunnel.io/m

▀▀▀▀▀▀▀ ▀ ▀▀▀▀▀▀▀
▀ ▀ ▀▀ ▀ ▀
▀ ▀▀▀ ▀ █▄ ▀ ▀▀▀ ▀
▀▀▀▀▀▀▀ ▀ ▀ ▀▀▀▀▀▀▀

Private build first. Your local AI workspace next.

Sign Up for Early Access Browse Docs