Setup Wizard
The Setup Wizard is a guided, step-by-step flow that gets you from a brand-new account to a device that is running a model — all in one place, without leaving the dashboard.
The first time you sign in to a new account that has no devices yet, the wizard opens automatically and walks you through four steps:
- Name your device
- Platform — choose where the device will run
- Connect — install the agent and wait for the device to come online
- Deploy — deploy a recommended starter model
The Setup Wizard opens automatically once for new accounts that have no registered devices. After you sign in, it appears as an overlay on your dashboard. You can leave it at any time and come back later — see Pausing, resuming, and dismissing the wizard below. Prefer to set things up manually? You can always register a device and deploy a model from the Devices and Models pages instead.
Step 1 — Name your device
Give your device a friendly name so you can spot it later in your dashboard. The field defaults to my-device and is required. Loc.ai generates a one-time registration key for this device behind the scenes.
Enter a name and click Continue.
Step 2 — Choose your platform
Pick where your device will run. macOS is selected by default and is the quickest path — a one-click app download with no terminal commands.
- macOS (Apple Silicon) — download the Loc.ai:Link app and let its Setup Assistant register the device for you. This option carries Alpha and Apple Silicon badges to show its current scope.
- Advanced: install via terminal — expand this disclosure to install from the command line instead, with a card for Linux / macOS (terminal), Windows CMD, or Windows PowerShell.
On macOS, click Continue. If you picked a terminal platform, click Get install command.
The one-click Mac download is for Apple Silicon (M-series) Macs and is currently in Alpha. On an Intel Mac, open Advanced: install via terminal and use the Linux / macOS terminal card instead.
The platform picker is keyboard-navigable, and the wizard respects your system's reduced-motion preference.
Step 3 — Connect your device
How you connect depends on the platform you chose in Step 2.
macOS — download the app
Click Download for Mac to download the Loc.ai:Link .pkg installer. Open the downloaded file and follow the Setup Assistant: it signs you in, registers this Mac on your account, and starts the agent. There is no registration key to copy on the Mac path — the Setup Assistant handles registration for you.
For the full macOS walkthrough, see Install on macOS.
Install via terminal
If you chose a terminal platform (or opened Advanced: install via terminal on a Mac), the wizard shows a single, ready-to-copy install command tailored to the device name and platform you chose. Click Copy, then paste and run the command in the terminal of the device you are registering (not your local browser machine). Use Terminal on macOS or Linux, or Command Prompt or PowerShell on Windows. The terminal will prompt for your password.
The command installs the Loc.ai:Link agent, registers the device using its one-time key, and starts the agent.
On the terminal path, the install command includes a registration key that is valid for 24 hours. If it expires before you run it, restart the wizard (or register from the Devices page) to generate a fresh key. (The macOS app path has no key to expire.)
Waiting to connect
Whichever path you use, the wizard waits in the background and detects your device automatically — there is nothing to refresh. Once the agent comes online, the status changes to connected and ready to serve models.
On the terminal path, the device must register under the same name you entered in Step 1 for the wizard to recognise it. On the macOS app path, the Setup Assistant names the device, and the wizard detects the newly connected Mac automatically.
When you see the success message, click Continue to deploy.
Step 4 — Deploy a model
The wizard recommends a lightweight starter model so you can see your device run a model end to end. Click Deploy model to deploy the recommended model, or choose Browse other models to pick a different one.
A progress bar tracks the deployment as the model downloads and is configured on your device.
When the model finishes deploying and starts serving, the step ends on a model is live state. From here you can select View it under Serving to open the model in the Serving view, then click Done to close the wizard.
Choose Skip for now, deploy later to finish setup without deploying. You can deploy a model at any time from the Models page.
Pausing, resuming, and dismissing the wizard
There are two distinct ways to leave the wizard:
- Minimise it — clicking the X in the header (or clicking outside the wizard) minimises it to a Finish setup — N of 4 steps button in the bottom-right corner, with a small progress ring. Click that button at any time to pick up exactly where you left off.
- Hide it for good — clicking Not interested opens a confirmation dialog. Confirming permanently hides the wizard, and it stays hidden across browsers and sessions.
Whichever you choose, you can always add devices from the Devices page and deploy models from the Models page as normal.
Once your device is connected and a model is deployed, continue to Run First Inference to send input to your model and view results.