MidJourney Inspirational Prompts

AI-Generating imagery is here to stay, so here are some of my favorite prompts that can be created using MidJourney which is now in open beta on Discord.
Here are all of the prompts from each series if you'd like to play around with them yourself!
PROMPT 1
golden hallway of the heavens, giant marble pillars, white granite walls, onyx floors, mist, fog, volumetric light, city of light, realistic, hyper realistic, depth of field, wide aerial frame, --ar 16:9
PROMPT 2
city of white marble with roman columns in the sky :: landscape :: greek gods :: Zeus :: Golden light :: mist :: photo realistic :: 8k :: --ar 16:9
PROMPT 3
Human Skull wrapped in vines + black paper + elements + red + gold + neon + baroque + rococo + white + orange+ ink + tarot card with ornate border frame + marc Simonetti, paul pope, peter mohrbacher, detailed, intricate ink illustration --ar 9:16
PROMPT 4
::Volcano erupting at night :: mountains, night sky 8K, stars, flames, smoke ::volumetric light:: Atmosphere : ultra-detailed, realistic, landscape, hyper-realistic, wide --ar 16:9
PROMPT 5
::jungle landscape :: snake river, vines, oil ruines, sunset sky 8K :: volumetric light:: Atmosphere :ink + ultra detailed, intricate digital illustration --ar 16:9
PROMPT 6
::interior:: Prometheus spaceship cockpit from 2455, dull lightning, circuit boards, textures, computer components, volumetric lighting particles, debris, Captians chair, super detailed, high resolution, 8k, octane render, --ar 16:9
PROMPT 7
cinematic environment with very high, tall, sharp, red and blue mineral crystals, everywhere, tower of crystals, High Angle Shot, gloomy,
nebula cloud, looking away, high details, cinematic, Octane Render, 8k, --ar 16:9
PROMPT 8
brave new world, vray, architecture, war is peace, slavery is freedom, blue accent, atmospheric, backlit, epic, cinematic, detailed, wide angle, concept art, --ar 16:9
PROMPT 9
Spaceship in city blade runner 2049 futuristic sci-fi, dark, rainy Dark storm, bright lighting, contrast, Raging, terrifying
PROMPT 10
::frozen tundra landscape aurora borealis in the sky :: mountains, twilight sky 8K :: volumetric light:: Atmosphere :ink + ultra detailed, intricate digital illustration --ar 16:9
PROMPT 11
St. Patrick's Cathedral with large stained glass windows :: Grand hall, volumetric light :: light rays shine through the glass
broken pews | clean :: cinematic :: hyper realistic :: 14mm, wide angle --ar 16:9
#midjourney #ai #district7

[@phillyguy199]

Thank you so much for sharing the prompts. These are all amazing!

[@profsteve7653]

keywords like cubist and cubo-futurism with guitar, mandolin,face,piano,sky scrapers,city,buildings ,artistic,textured 3d gave a collection of absolutely stunning realistic concepts of that period easily of frame able quality-this is remarkable

[@District7DrewGeraci]

Thanks for sharing!!

[@Soul_809]

I think A.I creates a different image even if you type the very same prompts,It means once an image is created by a user no one else will create the very same image ,& I love that thing

[@District7DrewGeraci]

Yes! This is 99.9% true. If you use a 'seed' setting you can achieve a nearly identical image twice, but if no one knows the seed number then it will always be unique.

[@marianaz2015]

absolutely stunning , very creative prompts

[@pawnoir]

Nice pictures, good editing, smooth voice 👍Thank you

[@LeandroT242]

Prompt Mania is amazing, thank you for sharing this =D

[@CarolineJensen]

Drew, I’m totally obsessed with this now!

[@milovansluis]

I like how you say its ur favourite with almost every prompt (really wholesome)

[@District7DrewGeraci]

They're all my favorite, hahaha

[@samet8367]

This was just awesome! Thanks a lot!

[@District7DrewGeraci]

Thanks for watching!

[@danielgollner1390]

Why do your videos not have millions of views? It's so beautiful, the music, your voice and explaining what you did even giving us your prompts

[@District7DrewGeraci]

That's incredibly lovely of you to say and I deeply appreciate it - hopefully one day that will be the case! Until then, thank you so much for watching!!

[@masihnewbie0]

Thanks for the inspiration! One think that im notice is A.I. is really good when it creates object image, but sometimes it's kind a confused when creating subject image like human (usually have 1 or 3 leg, or additional face in it chest XD)

[@thomashovgaard3134]

mindblowing

[@IAMTHESWORDtheLAMBHASDIED]

OMERGGGERDDD M808 THANK YOU FOR PROMPT MANIA OHHH wowowowowowow was damn near finished with one mindmap//uml//diagramming such that I could spend less time giggling at the cute little realistic//pixar'esque animals and moody paintings etc and upscaling into infinity lololol, IOW I was going to pen and paper this site > >>>>> than anything I would've likely ended up giving up on LOL thank you so much it's stupendous!

[@District7DrewGeraci]

Woot! You're very welcome and thanks for watching!

[@victorflores5333]

It’s okay to feel that why. I felt like that sowhat at the beginning. Just use the love you have for making soft to overpower that feeling

[@tuanvumai208]

Thanks bro..

[@massimilianocesaretti4988]

Great video! Which "Style" paramether did you use with these prompts? Low, Med, High or Very high?

[@seasidevfx3109]

Very nice -

[@District7DrewGeraci]

Thanks for watching!

[@taniavolskaya4904]

Thanks for awesome video. Could you share with us what was is the prompt for your cover image to this video?

[@vishalpatel1107]

You mentioned making a print from your mid journey creations, what has your experience been thus far with regards to resolution and size of the print?

[@aaronbazil]

Man, I just really wanna learn how to paint like that irl

[@winstoneverlast]

Beautiful work and very helpful! I need to expand my use of variables and weights to maximize the system. Were you using version 3 for these?

[@District7DrewGeraci]

Thanks for watching! This was done in version 2. Version 3 really changes a lot of things and honestly is way easier to use/manipulate.

[@swapnilchhabilwad6540]

@@District7DrewGeraci Hey .... Drew Geraci ... could you tell the name of background music ....

[@FilmFactry]

I want to learn more how best to include REFERENCE images, and multiple image to my prompt to steer the render.

[@iam607]

Man I've been uploading references and the results are just so different. Even when I put in --1 w. I think they're avoiding people doing derivative copyright infringing stuff. I gave up lol

[@Glacially]

The skull one is super cool! :O Also, how do you think of these really cool, creative and detailed prompts?

[@District7DrewGeraci]

Yeah, the skull one is pretty amazing! As far as thinking about prompts, I usually try to find other artists works (painters, photographers, etc) and ask the AI to generate the art in their style. So for the skull one, I basically told the AI to find other examples of tarot cards and then insert the other imagery that I'm thinking about, into that particular style. It doesn't work all the time, but I'd say it does 90% of the time. Thanks for watching!

[@dogmaticCURE]

Thank you for a great inspirational video. Many images are just so astonishingly artistic and beautiful. But I wonder what is the maximal resolution of images? I would love to use Midjourney to create some images that I could put on my wall. Is the resolution big enough to print 100x70cm images without losing quality?

[@District7DrewGeraci]

Thanks for watching! The max resolution is about 1920x1080p, which is at best, an 8x10" image. It'll be a few more years before we see higher resolutions for large format printing.

[@dogmaticCURE]

@@District7DrewGeraci Thank you for answering. I saw your video on my 75" 4K TV and the images you used in the video looked absolutely amazing. So I got the impression that the resolution was high enough for large print format. But I guess that images that look more like artistic paintings, and not detailed photographs, could still look good when printed in large format.

[@District7DrewGeraci]

@@dogmaticCURE I've upscaled the images to 4k so they 'should' look better than what I get out of the actual prompt generator, that being said, you could print these, you'd just have to upscale them to at least 8k for a decent large format print. It's not hard to do, but it'll require a couple hundred dollars worth of software to do it.

[@dogmaticCURE]

@@District7DrewGeraci That can be done in Photoshop?

[@zdenekburian1366]

@@dogmaticCURE better to use specialized upscaling sw, working with AI themselves, like topaz

[@AlexPlaksin]

Hey man! Stunning review, thanks! Did double dots help? (::)

[@District7DrewGeraci]

Thanks for watching! The "::" are weights, and usually help add detail to the scene.

[@AlexPlaksin]

@@District7DrewGeraci got it, thanks!

[@hypersonicmonkeybrains3418]

These can be upscaled further for more detail and sharpness, but they need a very good transformer model upscale not the usual crap you find on the web. Try swinIR on replicate*. But you might have to break the image into tiles and stitch them back, because that upscale model needs a lot of vram.

[@jessedoyle4308]

I've been curious about getting different skin complexions. Like human with jellyfish skin or celestial skin, where stars would be freckles on a universe like skin. That barley makes sense to me haha but I'd love to see what ai does with those kinds of prompts, I feel there are some examples in games like elden ring, with the way they depicted some of there demi gods. Im going to love this when I get my hands on it!

[@peterson61]

Hi and thanks for the video. Please could you explain the use of the :: ? To date I've seen them placed/used as frequently as those shown. Many thanks...

[@District7DrewGeraci]

Yes, the semi-colons are used as weights. They put focus on certain subjects of your shot. You can assign numerical values to increase or decrease their importance in the scene. They aren't as much of a thing now with V4 (since you don't really need them) but I hope that helped to answer your question.

[@peterson61]

@@District7DrewGeraci many thanks as I thought they had to be followed by a weighting

[@District7DrewGeraci]

@@peterson61 you can add numbers (0 to 2 I believe to add specific weight, where a :: 1.1 would outweigh a 1.0 etc.

[@PhylomedusaOpasna]

Please help with a question. - Can I use the works that I made in this neural network for the cover of a music album or book cover? (It seems to be impossible for free, but if through a subscription?) - and the second question, is it possible to use them for the same purposes if I change the detail in Photoshop? (Are small changes enough, or are big changes needed?)

[@cotoba]

about the first question u can use for commercial purpouse if you subscribe.

[@samjones9077]

Yes as long as you subscribe, unless you make more than 1m dollars annually in which case it's still okay as long as you also buy corporate subscription

[@ebutuoyjg]

hi, were these from version 2? or was it beginning of 3? thanks. nice prompts

[@District7DrewGeraci]

This was from version 2, a long while ago, ha.

[@RipsterBolton]

Hi! Nice video! How much you pay for the subscription of MidJourney?

[@District7DrewGeraci]

Thanks! It's either $10/200 images or $30/unlimited mo. - I am not sponsored or affiliated with Midjourney either.

[@MaxWeir]

Are you planning on doing more of these?

[@District7DrewGeraci]

Yeah - it's been a busy couple of months! Need to post more :)

[@cicadashroud2531]

For the 30/mo option, can you upscale to full resolution for print?

[@District7DrewGeraci]

It depends on how large you want to print. Most of the images are less than HD (1920x1080) so you can (at best) make a 5x7 that looks good. You'd have to scale up the image (via another AI service) to print any larger.

[@Nutronic]

Do the symbols such as :: , and + make any difference or is that just for you?

[@Nutronic]

@@PsychoPlacid Ah interesting, Thanks. Might give it a try.

[@BigPromise]

Quick question, what is the difference between using commas and :: between words(if any)

[@District7DrewGeraci]

Nothing anymore. It was really only meaningful for V1-2 of Midjourney. Colons are weights (for the subject) and commas allow for additional subjects or elements.

[@RBYW1234]

1:34 - so i think i seen this - you know how your eyes make up like 40% of what we actually see in our brain. When you close your eyes, you still do this, but kind of with imagination ./ dreaming. Anyway, i think i seen this like last week...

[@adamstevenson7534]

You use colons and commas in your prompts, and I've seen quotation marks and hypens used too. What difference do they make?

[@District7DrewGeraci]

It compartmentalizes certain aspects of the image and tells the AI what to prioritize. Adding weight(s) has a similar effect. I've found that the colons work the best for specific subjects. Commas have less weight and those characteristics are added to the scene more loosely. With the new V3 that just launched it makes all of them kind of the same now, so there's no real difference.

[@SerenityDreaming]

@@District7DrewGeraci this is excellent information. One of the more helpful channels I've interacted with in a long time

[@District7DrewGeraci]

@@SerenityDreaming that's awesome to hear and I sincerely appreciate that!

[@michellcastro1970]

To download soft soft is the download free or do u have to pay for it?

[@District7DrewGeraci]

MidJourney is $30/mo for unlimited use or $10/mo for around 200 images.

[@locomotionchannel]

I've tried to look for the :: in the documentation but can't seem to find it, why use :: ?

[@skyko]

You use the :: after a prompt to give it weight (of importance) example-- Bowl of fruit::2 house fly::.5

[@luxurycardstore]

You have some awesome prompts, but the music is too loud making this video difficult to hear what you are saying.

[@nathanstone6838]

What's the promt website link.?

[@District7DrewGeraci]

Promptomania.com

[@DonPino596]

Can I run this on a 16gb amd gpu?

[@user-lp3yp3pb7e]

You can even run it on your phone.

[@xapes1]

how do you make prints of them?

[@District7DrewGeraci]

Right now you can upscale the image to a 2k x 2k resolution so you can likely print out (at the most) an 8x10 or if you AI-scale ip using other programs maybe 16x20 but it'll be pixelated.

[@District7DrewGeraci]

I've just released a new video on printing if you're interested in watching

[@bloodswarms]

I've looked at all of these AI image generation tools. The fact that none of them produce images of a printable resolution kills it for me.

[@District7DrewGeraci]

They're definitely printable, you just have to AI scale them up (in Topaz or Lightroom) to be able to print anything between 8x10 and 16x20.

[@bloodswarms]

@@District7DrewGeraci I would love to see a video of that process. I tried Topaz and other methods and couldn't find a way to make the image usable in a quality print. But I also don't want an 8x10. I was wanting to print at typical wall art print sizes, which I think would be the only reason to print them. 24x* or 34x* or larger. I like what MidJourney and other AI prompt-based generation tools are capable of in terms of visuals for screen, but as far as printing for display it seems useless.

[@District7DrewGeraci]

@@bloodswarms I don't think you'll ever see any IA generator producing resolutions that could print at those ranges simply because it would take too much computing power to produce, which would be a huge strain on servers. Until the process can be sped up and expanded we likely won't see increases in resolution for a while.

[@Andonchoto]

Maan. You got to do the painting on em ..

[@nareshbc1042]

How do you change size of image isn't it is that it generates square size image

[@baruc4258]

just type -- ar 16:9 (16:9 is the ratio, it can be 9:16, 24:9, etc..)

[@nareshbc1042]

@@baruc4258 can it be used for commerical purpose like printing and other stuff

[@soumyamty]

@@baruc4258 oooo thank you I was wondering the same

[@LeandroT242]

@@nareshbc1042 yes u can use but you need to create those images in subs mode, if you use the free u need to credit MidJourney. See the FAQ in the website

[@scottwarner21]

who owns the artwork copyright?

[@cosmosapien3434]

no-one its there

[@District7DrewGeraci]

Technically YOU own the copyright, but it also gets a bit grey when you incorporate other people's work. It's a new frontier for copyright, that's for sure!

[@edsonlucascardozo3226]

What was the prompt for the tarot cards?

[@jomontanee]

Just “portrait” it.

[@StreetRallyAsh]

+ tarot card with ornate border frame

[@District7DrewGeraci]

Human Skull wrapped in vines + black paper + elements + red + gold + neon + baroque + rococo + white + orange+ ink + tarot card with ornate border frame + marc Simonetti, paul pope, peter mohrbacher, detailed, intricate ink illustration --ar

[@andrewmartin8565]

Can one create artificial images of celebrities?

[@District7DrewGeraci]

Yup, or at least, get pretty close to it.

[@andrewmartin8565]

Thank you for the answer! Yout images are awesome.@@District7DrewGeraci​

[@soumyamty]

Umm... I just tried out midjourney today on discord but I’m new to discord, so can anyone tell me is there a way I can see only my own messages and not the entire chat because that moves really fast and I had to keep scrolling the whole time to find my prompt

[@District7DrewGeraci]

Yes, if you're a $30/mo subscriber you can just DM the MJ Bot and write your prompts there. If you are a free user or $10/mo subscriber you can only use the channels provided which can be cumbersome. The $30/sub is the best option.

[@soumyamty]

@@District7DrewGeraci oh thank you so much for letting me know! So does that mean in the $30/month account, the DM prompts are really private and others can’t see my prompts/generation and cannot interact with it or are the generated artworks still public and can be interacted with like create variations, etc. by other people?

[@District7DrewGeraci]

@@soumyamty The only way to do that is to start your own discord server and invite the midjourney bot to it so you can create there. OR you can pay an extra $20/mo on the midjourney server to have them make the prompts private for you.

[@soumyamty]

@@District7DrewGeraci Okay thank you!

[@RBYW1234]

@@District7DrewGeraci Do we get to call this content ours? or is it property of the A.I? Do we share credit / ownership?

[@thisisres]

I tried a prompt but nothing happened... I literally even copied one of yours... still nothing. Should be on trial as account is

[@District7DrewGeraci]

Are you in the discord channel and using /prompt ? Servers are up so there shouldn't be any issues making art.

[@deaddropholiday]

Why does so much of MidJourneys stuff look the same?

[@soumyamty]

Because the AI has developed its own style lolll OR it was fed only a certain style of images

[@deaddropholiday]

@@soumyamty Yeah - but I'm seeing the same style from different operators. Me thinks something fishy here ...

[@soumyamty]

@@deaddropholiday Yeah LOLLL freaking AI creeping us out... I feel like the AI is trying to harvest us and will turn on us someday hahahaaa eek

[@soumyamty]

@@deaddropholiday which other operators have you tried out?

[@deaddropholiday]

@@soumyamty I'm just going off sample images mostly pulled from Google image search. So I'm guessing they aren't all from the same operator. Would be extremely co-incidental if they were. But there's seems a lot of commonality in style. Unless everyone is just copying everyone else's parameters. Weird.

[@orlandofabiansanchezgarfia767]

The slower you put the video the more northern you soft ahahahahah

[@alltheserobotsshallfall]

you have a couple of bad words on the "future" prompts, no wonder how the future sucks in there - war, slavery - could not go well

[@caerbenogg]

AI art is not art! If you use the generated image as a basis and modify it yourself, mix it with your own created images, than it could be your own art. People are selling NFTs which are 100% MidJourney created and they think they are artists.

[@SerenityDreaming]

What is the difference between art we make with this tool and art we make with other tools?

[@caerbenogg]

@@SerenityDreaming Hi Kimberly, the difference is: you don't make it, you type in words and the image is generated. You had zero input skillwise and the result is outside your influence. Art without human emotion and physical craftmanship is not art, because it became dehumanized. It's not a tool like Photoshop. In Photoshop you need some skills to produce an image and do compositing, the tool in itself is not creating the final image. AI is different, it does all "creative" work for you and you can't control the output. The result is more or less a coincidence. Selling 100% AI generated images is morally questionable. If everything can be art, nothing is art anymore.

[@RogersPhotographyGuilford]

@@caerbenogg After using Midjourney for a couple of days now, it's quite apparent that it takes a tremendous amount of time and some "skill"/knowledge to string together the correct words/phrases to achieve an image you are in search of. Is the output art? Yes it is. Who are you or anybody to define what art is? Is a painted solid black square art? Is a painted solid white square art? Is a reproduction of a soup can art? In the not so distant past (which I can clearly remember) photography was not considered an "art" discipline. Today it's just as "valuable" as the hand painted/drawn image. Photoshop was just a "toy" and look at it now. This is just another tool which you can leverage to enhance your vision(s). You need to stop being so angry and embrace and use new technology to your advantage. The base of art is technology. You're not going to change the tide and this is certainly a tide. It engages the mind visual senses (part of the mind I guess). Be happy!😁

[@marvinr.rissmann5248]

I don't know if I understood the comment correctly, but I agree with you.everybody is hyped for midjourney and hey it is cool but does it still have something to do with art? especially the craft itself? For me it is sooo dangerous. Interesting for creating and idea how could it Looks like but it shouldnt be more. The fact is: it is a AI. Just because you Tipp some words you will Never have the Same level of professionelism an self thought artist has.

[@caerbenogg]

@@marvinr.rissmann5248 100%

[@wojtekgol4658]

Ok, your " tutorial" is how to create images by using words! It's very "inspirational"! You didn't create those images! And you will never be able, because your skills are random "hit or miss" words! Ok, try to do useful art with what I wrote. And I mean in your brain not in a software, where can anyone be an artist as soon as he or she can say a few words .

[@cotoba]

lol we find a frustrated artist who doesn't sell his art and is sad because he's going to be replaced by Artificial Intelligence So funny

[@Jimmy_The_Goat]

incredibly cringe