olace

[ get started ]

Quickstart

Install Olace on a computer, run a model on it, then pair your phone so that model is in your pocket, at home or a world away.

1. Install Olace

Download Olace for Windows, macOS, or Linux from olace.app/download.

Setting up a home server or a rented GPU box with no display? The install script sets up the standalone daemon and CLI, and every step below works the same from the shell. See Headless & cloud hosts.

curl -fsSL https://olace.app/install.sh | sh

2. Set up local AI

Open Olace and choose Set up local AI. Olace detects your hardware, installs a local runtime (Ollama, LM Studio, or llama.cpp), and recommends a starter model that fits your GPU.

The same flow is available any time under Settings › Local AI, or from the terminal with olace setup.

3. Start a chat

Type a message and send it. The model runs entirely on your machine, so the first response may take a moment while it loads.

Explain why the sky is blue in one paragraph.

4. Sign in

Signing in is passwordless: a one-time code by email. It unlocks the parts of Olace that no local runtime gives you on its own: Remote Access to this computer's models, and zero-knowledge conversation sync across your devices.

5. Pair your devices

The payoff. On the computer, open Settings › Remote Access and show the pairing code. Scan it with your phone, or claim it from another computer with olace pair claim <CODE>. Done: this machine's models now work from that device over an end-to-end encrypted connection, on your couch or on another continent. No VPN, no port forwarding, no setup.

Any of your devices can be the consumer: your phone, a laptop using your workstation's GPU, a second computer. The host can be any computer with a GPU: the one on your desk, a home server, or a GPU instance you rent. Hosts with no display are set up from the shell and pair the same way, with Headless & cloud hosts.

Close the laptop mid-conversation and pick it up on your phone; with sync on, every device stays on the same page.

Next steps