Prompt Design Resources for MidJourney AI

by Shane McGeehan

Blog: Recent Entries

Newest Posts

Happy Amoebas

MidJourney’s Facebook is doing weekly themes for fun, and I love fun, so let’s do it. This week’s theme is “creatures” so I wanted to think of something out of the ordinary. I own a pretty powerful microscope, not for the sake of science, but purely aesthetics. I hook up my camera to explore these tiny worlds and make photographs. Anyway, I was wondering, what if you looked in and saw a bunch of happy amoebas?!

I first made a bunch of happy amoebas with the prompt below, but swapping out --style 4b for 4a, or swapping between low and high stylize and chaos. I like running the same prompt idea in a bunch of ways to see what options I can work with. I then had the idea to make a bunch of fantasy microscopes to pair with the amoebas. I did the same with this prompt below, running it, and swapping numbers. Then I downloaded, upscaled, and altered some colors of the images, even using the new Photoshop Neural Filter to colorize some that were less colorful. And there you have it, a bunch of happy amoebas!

(click the prompt to copy it!)

/imagine prompt: microscope imagery of a happy amoeba smiling and laughing in a petri dish surrounded by other cells --ar 3:2 --style 4b --chaos 0 --stylize 1000 --v 4

/imagine prompt: product photograph of a high powered microscope, close up, lens, slide, bright light --stylize 800 --chaos 90 --v 4 --style 4b

(click to view)

Photoshop Pattern Actions

Want to try out some free Photoshop actions? Read on! - I can remember being a kid staring at a big old computer monitor and making mirrored images using MS Paint. That’s how far back my passion for repetition goes. More recently, I began developing a collection of Actions in Photoshop to speed up pattern-making. I used Adobe Illustrator to create the perfect size shapes needed, then imported them into Photoshop. A lot of steps are involved (cutting out triangles, rotating them perfectly, filling gaps), but once you do it right and record it, then you can run the Action with little effort. So far I have created over 50 actions that do all kinds of things to an image, and when you combine them, things get really wacky.

Click HERE to download a sample pack of my actions. Just open an image, click “play” on an action, move the crop to where you want and hit enter - it does the rest for you!

For the example image here (this is less art and more just a sample of what these actions can do), I came up with a really long and weird prompt, as you can see below. I wanted something abstract yet detailed to work with. Upscaled a bunch then brought them into Photoshop. Just to show off what the actions can do, I ran each image through a few different pattern Actions - ending with a hexagonal one so I could arrange them in a grid. Kinda trippy, dude.

(click the prompt to copy it!)

/imagine prompt: close-up photography of molecules and atoms twisting into dna hallucinations and crystals, insanely detailed and complex, fractal sacred geometry, dmt, volumetric lighting --no defocus --stylize 1000 --ar 3:2 --v 4

(click to view)

New Technologies GIF

As creatives, sometimes we get “stuck” and aren’t sure what to make, but it’s best to just keep making. I challenged myself to create an animated GIF in less than an hour. Much of what I do is from “play” like this, experimentation, but bigger work usually grows from it.

The foreground was made using /blend on two MidJourney images using the prompt below (with some terms and style swapping). I processed the images in Photoshop, created masks, and exported PNGs with alpha transparency. I then used Premiere to add rotation to the interior as well as an animated “searchlight” look in Premiere. I imagine this as some kind of advanced technology, but honestly, I don’t know what it does… yet.

(click the prompt to copy it!)

/imagine prompt: A portal, metal circuit boards, wires and gears, led lights hue glow insanely detailed stone and rock circular door --style 4a --v 4

(click to view)

USA

This is a piece I began a little while ago and finally felt the need to finish, given the recent news. I don’t plan to get political on this page often, but I also don’t care if I turn people away because of this particular frustration. Something needs to be done, and we never do enough.

Since I use this blog to talk about techniques, this one involved a few. I created large JPGs of separate letters and then used them as image prompts in order to make the stylized font (a method I came up with for getting cool fonts in MidJourney). The prompt for the letters are linked below. It took some random weight changes to get the look I wanted. I then upscaled in Visions Of Chaos, cleaned, and composited in Photoshop. Lastly, I made an image of the flag covered in ammo, reversed it (a signal of distress), and added final touches like masking, shadows, and painting slight tones of color in.

(click the prompt to copy it!)

/imagine prompt: IMAGEURL disassembled ar-15 rugar assault rifle::10 dripping red paint in background::7 --ar 2:3 --v 4

(click to view)

CUBE³

I began this project a while ago (back in version 3) but set it aside while working on new ideas. I decided to dig up all the footage and finally get the idea out. I wanted this to be an experimental and experiential piece. Fast and confusing, and somewhat overwhelming… I ran the same prompt over and over (an image prompt from another image I made in MidJourney), downloaded all the video files, then upscaled the video. This better resolution allowed me to crop each of the grids into four individual videos. The audio is from an older work where I mixed various recorded sounds in Adobe Audition. Overall I am quite happy with the results, but would love to see this projected on a wall as a final form.

Pro Tips: I learned a few useful things working on this project. (1) On V3, if you use the --video parameter, you usually get a short little video, but if you pair it with --quality 5 as well, you get over 30 seconds of footage! I am hoping they add this video feature into V4, or maybe V5 when that comes out. (2) Since the video output is only the original grid, you can AI upscale the footage to get a larger resolution file, then crop in to one of the images in the grid. I used Visions Of Chaos to upscale but there are other options out there.

(click the prompt to copy it!)

/imagine prompt: https://s.mj.run/1AxynIXvqFs --iw 2 photograph of a glass cube with dripping black liquid::3 200mm lens, extreme closeup:: volumetric lighting:: disco color:: --no people, faces --video --stylize 1000 --quality 5 --chaos 90 --ar 5:4

(click to view)

Sunsets On Other Planets

Sometimes I make weird things, say I’m going to make more, then move on to something else and forget - only to find them over a year later… The image on the left was sculpted in Virtual Reality, textured, lighted, rendered in Marmoset, then composited with my photography.

On the right is MidJourney’s interpretation. I used an image prompt plus “sunsets on other planets” (which is the title for the still-in-progress series). I later went further by Remixing and adding to the scene. A thunderstorm, asteroids, fire, fog, snow.

MidJourney is a fun experience in creating. It’s exciting, like a good hook at the end of a television show and you have to keep going. That said, there was something awesome about listening to music in VR and sculpting a giant alien landscape. Similar to AI, I had no idea what the final image would look like, but just went with it. In summary, I need to get back into VR more too.

Below is my full image prompt. Image link and everything. Just click-to-copy then travel to my planet! Add things, change your perspective. Make it what you want!

(click the prompt to copy it!)

/imagine prompt: https://s.mj.run/5NpdzBwYUH0 sunsets on other planets --ar 3:2 --v 4

(click to view)

ChatGPT Learns MidJourney

ChatGPT is the newest talk of the town in AI. The old Version 3 was neat, but the latest 3.5 is just so smart that it’s creepy. GPT, or Generative Pre-trained Transformer, is a text-to-text AI that attempts to respond as if you are having a conversation with another person. It can be used to chat with, write stories, poems, or even code a script for Photoshop. Yes, seriously. Like other AI models, it is trained and cannot take in more information, however, during a single chat session it can remember what you tell it.

Months ago I used the old version to assist with MidJourney prompt writing, but I thought if the new version can remember things I tell it, maybe I can teach it how MidJourney works? I then wrote out how prompting works, format, algorithms, aspect ratios, and even what parameter settings work with different versions… and you know what? It worked!

All I do is copy and paste my “lesson”, hit the voice-to-text button and speak what kind of prompt I am looking for. After it gives you a result you can even ask it to replace things, change the version, or alter the prompt text. I foresee prompting changing in many ways, but one clear need is a voice-to-image interface - and something like this might be the way… Now, it’s not perfect. It has accidentally given me --stylize 5000 on --v 4 which isn’t allowed, but this is just a small test, and I bet if GPT gets to version 4 or 5 and can be actually trained on all of MidJourney - we will have an app on allowing us to make images by just talking to our phone.

(click the prompt to copy it!)

/imagine prompt: a futuristic city once bustling with life is now being overtaken by nature. Vines and plants grow through the cracks in the concrete, and trees tower above the crumbling skyscrapers. The setting sun casts a warm glow over the scene, creating a peaceful contrast to the abandoned city. --v 4 --style 4a --chaos 20 --stylize 70 --seed 420 --ar 3:2

(click to view)

Cursed

You wanna know a secret? I tried to keep quiet about this for weeks, but since someone else let the “easter egg” out, I figure I can share it here as well. The secret is… CURSED!

That’s right. Just add --style cursed to any V4 prompt (or remix and add it) and you get weird pixelated, textured, gritty, glitched, dark, and twisted images… It may not be for everyone, but it’s a super fun feature to play around with. I hope it sticks around. I honestly have no idea how it works, but I love using it!

(click the prompt to copy it!)

/imagine prompt: a dark dripping skull of an alien on display in a glass cube --style cursed --v 4

/imagine prompt: a kings throne made of metal pipes, a fancy rug below, and a stone wall in the background with black dripping paint, 4k, cinematic, wide angle, evil --style cursed --v 4

/imagine prompt: a whimsical sleep paralysis demon --style cursed --v 4

/imagine prompt: dried up angler fish human skull carcass on a sandy beach, giant teeth, bones, skin, scales, stormy skies volumetric lighting --style cursed --v 4

(click to view)

Publication: AI, Machines Rising

I was recently interviewed by Rob Wilson to offer some of my thoughts about AI and photography for an article in the digital publication of FRAMES Magazine… I also created the example images from the author’s photographs using 3 techniques and 3 different AIs - the attached is

It’s a worthwhile read if you are interested in new technology and how it might change the ways we work as image-makers. I was very honest and positive about how I foresee the two tools converging, as they already are and have been in the past… That said, I can’t help but laugh a little at how any time people talk about AI they act like the world is coming to an end. Of course it is, but chill, yo. (and don’t even get me started on Beno Saradzic’s out-of-context line about David Holz using images without consent, or how AI is breaking copyright law - that’s not true and not how it works)

Read the Article here:
LOOK CLOSER: “AI – Machines Rising ” – by Rob Wilson

(click to view)

Melting AI

I am a photo artist. Much of my work begins with a camera, but it isn’t your typical photography. I developed a process in Adobe Photoshop to “melt” my photos, and decided to try the effect on some MidJourney V4 images.

I won’t give away my magic trick secrets, but basically I create an Action that makes a bunch of little changes to an image, then I create another Action that runs the first Action hundreds and hundreds of times, using a Script to save out a frame with each step. Then when I compile the frames together in Adobe Premiere, you get this weird melting/dripping/trippiness effect… Just testing some ideas for a possible bigger project.

(click to view)

Blog Search

Blog Categories

Blog Archives

USA

This is a piece I began a little while ago and finally felt the need to finish, given the recent...

CUBE³

I began this project a while ago (back in version 3) but set it aside while working on new ideas....

Cursed

You wanna know a secret? I tried to keep quiet about this for weeks, but since someone else let...

Melting AI

I am a photo artist. Much of my work begins with a camera, but it isn't your typical photography....

Christmas Cards

I decided to make my own Christmas cards with MidJourney AI this year. To be honest, I rarely ever...

Mooooon

I just completed a really important study so you all don't have to... I was wondering what would...

Pin It on Pinterest

Share This