How to Use Ollama from Your Phone (iPhone and Android)
TL;DR: Using Ollama from your phone means Ollama keeps running on your computer, and your phone is where you chat. On the same Wi-Fi it is free with LM Mini, on iPhone and Android, and you can even download new Ollama models from the phone. Away from home is a Pro feature.
Last post was long chats. This one is for the Ollama people, who have been very patient while I kept saying LM Studio.
Ollama is the easy one, you install it, you pull a model, and it runs. It does not have a phone app of its own though, so the model sits on the computer, and the question is how to talk to it from the couch.
That is where LM Mini comes in, and I should say it again, I am not part of the Ollama team, I just built the app that sits on the other end. The phone becomes the chat, the computer does the thinking, and the model never leaves your house.
There is one setting to change before the phone can connect. When you first install Ollama, it only works on that one computer and ignores everything else on your Wi-Fi, including your phone. You switch that once, the docs show how for Mac, Windows and Linux, and from then on your phone can find it.
My favorite part is the one nobody expects. I can browse Ollama’s models from the phone and download a new one straight onto the computer, without going to the desk. I read about a model, I tap, and by the time I wander past the computer it is sitting there ready. That is free, the same as chatting on your own Wi-Fi.
If you use both, you will notice Ollama shows fewer settings for each model than LM Studio does. LM Studio gives you more technical control, Ollama gives you less to get wrong, and LM Mini is happy with either, so pick the one that matches how much you enjoy settings.
When you leave the house, the same Ollama models come with you through LM Mini Home on a Mac, or Connect on Windows and Linux. That part is Pro, and the remote access post has the whole story.
Everything else in this series sits on top of it the same way. A Persona, a voice, a group chat, none of them care which program is running the model.
Next post is files, asking your AI about a document without sending it anywhere. Until then, pull a model and say hello from the couch.
Get LM Mini: lmmini.com