Use Your Mac's Local AI From Anywhere on iPhone, iPad, or MacBook

VoxyAI is now on iPhone and iPad, and the apps are free. That alone is good news, but the feature we are most excited about is the one that connects everything you own. If you run VoxyAI on a powerful Mac at home, you can use that Mac's local AI models from anywhere in the world, right from your iPhone, iPad, or another Mac. No API keys, no subscriptions, no cloud middleman. Just your own hardware, working for you wherever you happen to be.

This is made possible by a feature called LLM Host. Here is how it works, why it matters, and how to set it up.

What LLM Host Does

LLM Host turns one Mac into a private AI server for your other devices. Pair an iPhone, an iPad, or another Mac, and that device can use the local models installed on the host Mac through Ollama. There is no API key to manage and no cloud round-trip. The connection is encrypted end to end, and only devices you have paired are allowed to connect.

In practice, that means the AI lives on the Mac you trust, in your home, and travels with you through whatever device is in your pocket or your bag. Your text goes straight from your phone to your Mac and back. Nothing routes through VoxyAI servers.

Why Run AI at Home and Chat From Anywhere

Large language models love memory. The bigger and more capable the model, the more RAM it wants. A Mac mini or Mac Studio sitting at home with plenty of memory can comfortably run models that a lightweight laptop simply cannot.

LLM Host lets you put the heavy model where the hardware is, then reach it from anywhere. Run a large model on the Mac with the most RAM at home, and chat with it from your iPhone on the train, your iPad at a coffee shop, or a slim MacBook Neo on the road. The small device in your hand does almost no work. It just sends your message to the big Mac and shows you the answer.

You get the best of both worlds: the power and privacy of a serious local model, and the convenience of using it from the device you actually have with you.

Three Ways to Connect

There are three ways for a device to reach your host Mac. Pick the one that matches your situation. You can switch modes anytime, and already-paired devices keep working.

LAN only. Both devices are on the same WiFi or Ethernet network. This is the simplest and fastest option, and it requires no router setup. Great for using your iPad around the house.
Tailscale. A secure private network that works from anywhere Tailscale runs. It is free for personal use, up to 100 devices, and it is the recommended way to connect from outside your home. This is the option to choose when you travel.
Port forward. Open a single port on your router for direct access from anywhere. This is the advanced option and requires manual router configuration, so use it only if you really know your way around your router.

A simple rule of thumb: LAN for home, Tailscale for travel, and Port forward only if you have a specific need and understand the tradeoffs.

Pairing a Device

Setting up a new device takes about a minute. First, on the host Mac, open VoxyAI Settings, select the LLM Host category, and toggle Enable LLM Host. The first time you do this, macOS will ask for network access. That is normal.

Then pair the device you want to use:

On the host Mac, with LLM Host enabled, tap "Pair a new device". A QR code appears.
On the device you want to pair (iPhone, iPad, or another Mac), open VoxyAI and choose AI Provider, then VoxyAI Mac (Remote), then Pair a Mac.
Scan the QR code with the device's camera. If the camera cannot reach the QR code, tap "Show details" beneath the QR on the host Mac and enter the values manually on the paired device.
Once paired, pick any model installed on the host Mac and start chatting from the paired device.

Pairings are stored on the host Mac only. You can revoke any device at any time from the Paired Devices list in LLM Host settings, and a device can also unpair itself from its own side.

Let Paired Devices Wake Your Mac

A paired device cannot reach VoxyAI while the host Mac is asleep. macOS has a built-in setting that solves this, so your Mac wakes up only when one of your own devices needs it.

Open System Settings, then Battery.
Tap Options.
Set "Wake for network access" to "Always".

With that turned on, an incoming request from any paired device will wake the host Mac, handle the request, and let it drift back to sleep on its own. There is no need to keep the display on. The LLM Host settings panel even has a shortcut to open the right pane for you.

Private and Secure by Design

Because this is your data moving between your own devices, security is built into every layer:

Every connection is encrypted end to end.
Only paired devices can connect. Anything else is turned away at the door.
A short list of safe endpoints is allowed through, and everything else is blocked.
Pairing details and revocation are managed entirely on the host Mac.
Nothing routes through VoxyAI servers. It is a direct line between your devices.

One note of caution: Port forward mode opens your Mac to the internet. Every connection still uses HTTPS and only paired devices can connect, but any router-side misconfiguration is yours to manage. When in doubt, use LAN or Tailscale.

The iPhone and iPad Apps Are Free

VoxyAI for iPhone and iPad is free to download and use. Pair it with a Mac running LLM Host and you have a private AI assistant powered by your own hardware, reachable from anywhere, with no per-request fees and no data leaving your devices.

If you already run a local model on your Mac with Ollama, you are most of the way there. Enable LLM Host, pair your phone, and the AI you trust at home is suddenly in your pocket too. To learn more about setting up local models, see our guide on using Ollama as a free local AI on Mac, or open the LLM Host section of the documentation for the full details.