Substack's Free AI Image Generator

It's a bit rudimentary, but there appear to be no usage limits. How to use it.

If you want to generate some AI images…for Substack or elsewhere…you may not realize you have a free AI generation tool right under your nose here at Substack.

You can find it when creating an article and adding an image. Click on the image icon in the top toolbar, and you’ll see a little magic wand icon with “Generate” next to it.

The following prompt will appear:

As you can see, you have a number of “styles” available to go with your prompt. I have yet to play with all of them, but I find the Anime style to be pretty good, while “Comic” usually (not always) comes up with stuff that’s overly busy and chaotic. Each generation will result in 4 images. It’s not instant…you might need to wait a few minutes or even try again as it hangs sometimes.

Here are the results for “cyberpunk tinkerbell with magic fairy dust” and the Anime option:

Let’s see how it does with something photorealistic. Hmm, what should I prompt? I’ll try - “cow in front of a barn with storm coming, photorealistic” and the Photo option.

When I click “insert” on the first result, here’s what Substack does…it inserts the photo and the prompt as a caption:

cow in front of a barn with storm coming, photorealistic

(See below for more cows.)

Unfortunately, once I click “insert,” Substack pops out of the image generation window, and I lose the other photos. In fact, if you click outside the window by accident, you’ll lose all your generated images. This already happened to me once.

If you’ve played around with AI images like I have, you’ll know that it’s often hard to get exactly what you want, and if you see something that you like, you’d better save it, as it’s a once in a lifetime deal.

Therefore, you should always right-click every image you like that comes up and save it to your computer before doing anything else.

So what can you do if you accidentally click away and can’t access the other images generated by Substack? Well, if you are on a PC, and you want to bother with this level of obsessiveness, you can do what I did…install a cache viewer, which will enable you to save the images you just lost. This one is free: ChromeCacheView (scroll down the page for the download link).

Install ChromeCacheView and open the application. First, go to File and “Select Cache Viewer” or press F9 to select your browser. It should automatically find your cache files. Once your cache files are loaded. you should see 4 similar entries from substack-post-media.s3.amazonaws.com — these are your images even though they have no file extension.

Right click on each link and select “Open Selected Cache File With” or press Ctrl-W and use an image app like Photo Gallery to view and save the photo with a human readable name and proper .jpg extension.

See, I have rescued the other cows:

Unfortunately, that’s not really what I wanted. Most have no barn! The AI took “in front of a barn” literally; we must be looking out from the barn!

I’ll have to fiddle around with my prompts to try to coax the AI to pull up the image I want, which is the real art to using AI tools like this. Here’s my second attempt:

cow standing next to an old-fashioned red barn with a storm coming in the background, ominous clouds, winds, photorealistic

cow standing next to an old-fashioned red barn with a storm coming in the background, ominous clouds, winds, photorealistic

Nice sky but what the hell kind of barn is that? A red shed? Here are the others:

A headless cow? I’m going back to Anime:

cow standing next to an old-fashioned red barn with a storm coming in the background, ominous clouds, winds, photorealistic

That one is OK, but the Anime preset also gave me a version with a giant horse-cow:

The “Epic” setting wasn’t much better:

cow standing next to an old-fashioned red barn with a storm coming in the background, ominous clouds, winds, photorealistic

And here are some made with the “Paint” prompt:

cow standing next to an old-fashioned red barn with a storm coming in the background, ominous clouds, winds, photorealistic

Where’s the cow? Oh, she moved into the second painting:

This next one looks like it needs a Far Side caption:

Lily told Marge not to step into the Matter Transporter.

And here are two from the Comic option:

cow standing next to an old-fashioned red barn with a storm coming in the background, ominous clouds, winds, photorealistic

And the Toon version:

cow standing next to an old-fashioned red barn with a storm coming in the background, ominous clouds, winds, photorealistic

What’s that? A levitating bush on top of the barn?

By the way, the “Style” option is optional, so what happens if I use the prompt with no style? Well, I did label the image “photorealistic,” so let’s see what happens:

cow standing next to an old-fashioned red barn with a storm coming in the background, ominous clouds, winds, photorealistic

The last one is the best of this bunch, but the cow looks like a dog-faced cow. A Dog Cow.

Well, as you can see, using multiple settings and attempts, I didn’t quite get the image I was looking for, which is why human artists and photographers will still be needed for the foreseeable future.

However, if you are OK with an abstract and slightly wacky image to illustrate your Substack article, this free image generator is perfectly serviceable.