Image Generation with Local AI

TL;DR: Image generation in LM Mini lets you ask for a picture from the chat you are already in. AUTOMATIC1111 or ComfyUI on your computer makes it, and the picture comes back to your phone. A recent iPhone or iPad can also make one by itself, that part is still in beta.

Last post I taught the chat to look things up, and this one is for when I do not want words back at all.

I added it so I would not have to cross the house, unlock the PC, and find the right window, just for one sentence of prompt. The computer is already making the pictures and I’m making coffee, or I am already in bed, or just not in that chair. I wanted to type a prompt on the phone, tap once, and let the PC I already have do the work.

AUTOMATIC1111 and ComfyUI are programs on the computer that turn a sentence into a picture. LM Mini is not a new one, it sends the prompt to whichever of those I already have running.

I already had AUTOMATIC1111 for the days I just want a picture from a sentence, and ComfyUI when I already have a workflow sitting there. I type in the chat I am already in, and the picture comes back in that same chat, I do not open a second app to look at it.

Each picture is made fresh, the computer does not remember the last one. The chat does, though, so when I want a change I just say what to change, the model writes the new prompt, and I tap again.

I do not pay extra to do that at home. If the computer is already on, the phone can ask, and if it is not, my iPhone can still make a picture by itself with the free starter model (the bigger on-device models come with Pro). If I have already left, and I still want that computer making the picture, that is possible with Home on a Mac or Connect on Windows and Linux, a Pro feature.

If AUTOMATIC1111 or ComfyUI is already running on your computer, ask from wherever you actually are.

Next post is Memory, for all the things I am tired of telling a chat twice.

Get LM Mini: lmmini.com