I Tried Creating Art Using This Free AI Website.
Вставка
- Опубліковано 7 лют 2025
- Had some fun today playing around with this free AI website generating all kinds of characters and portraits. A must-use for concept artists?
Artbreeder link: www.artbreeder...
Thank you Iro Pagis for telling me about this!
"There's too much happiness for the human flesh, so it starts just dissolving"
in the words of a wise man "there is no happiness in life, only the spectre of it"
Hp love craft meets kafka.
Lol
This site is the perfect nightmare creature generator
I was about to say the same, I had to close the site because I felt uncomfortable with my unintentional nightmare creations staring into my soul.
@@alftuvik3820 boy out there scaring himself 😂😂😂😂😂
@@alftuvik3820 yep....
dog fruit. thats all i can say
Yes omg 😭 I got so many accidental horror demons widbeidn
Now I can slowly start mistrusting even more instagram accounts.
hahahahaha for reallllll
@@AshraafM I never got this attitude. Like, if someone figures out a way to make an image that never existed before, that everybody thinks is awesome... whats the problem? They still applied their taste and creativity (all we really do anyway, nothing is original) they didn't work hard enough? I mean sure working hard is admirable, but working smarter is... smarter lol
@@junovhs4646 I agree
@@junovhs4646 But if they're not telling their audience how they made the artwork, and let people assume they drew every detail themselves, that's a problem. You have to admit
@@turtletail313 so i guess there will come a time where arts are branded and classified into:
Pure, AI Assisted, AI
where pure is pure human creation
AI assisted a mix of Human and AI
and AI just AI made.
The end is nigh! Repent sinners and be saved by the glory of AI.
LMAO JUST SAW YOUR ART THOUGHT GOT TO THIS VIDEO AND WAS THINKING HMM SINIX SHOULD SEE THIS XD
Ey! Hi there Sinix, love you!!! :D)
So I shouldn't tell you guys about Nvidia Canvas then? :D
Пипец, Боро, Синикс комментит твои видео
*Rings bell* Hear ye! Hear ye!
I haven't had this much fun watching AI faces in my life - that shit got real macabre real quick
:D
True HAHAHAHA i
got dead, why?
But hey, all the art you draw is macabre, haha
7:15 too much happiness for the human flesh, it's dissolving
new scp
lol, that's genius))))
OD'd on Thalasin.
Oh yeah that's literally the SCP Mask lmao
This is useful. I no longer need real human models as inspiration to create concept characters.
What's worrying about this web application is that it can easily create monsters straight out of nightmares or horror fantasy.
Really? I need to try it. Im using Midjourney but still not good to create monsters just yet.
I’ve been using artbreeder for a bit over a year now. It has actually been really fun and useful for me as an aspiring writer/illustrator. I get really frustrated trying to find models and actors that look like my characters to use for photo references when I draw them, so I make them using artbreeder and I always have a headshot of my character to refer back to. I recommend it for anyone who want images of their OC that can’t draw or if you need references when no one looks like your character.
nice comment!
Its interesting because I also used to feel very frustrated due to the lack of constant or proper access to models to represent not only my character design, but also certain poses and movements that I was unable to find somewhere else. Due to that I started to understand why Disney used to reuse references in the past, but I am sure they still do, and every animation company still do, it is just that it is no longer that obvious. Now, one thing that I have learned and I am experimenting is this: let's say that you have already designed the character, you have the head already designed from 2 or 3 sides, then the body, front back, and thats it. Then, I see a movie or an animated series and I find a movement, pose or expression I want to use, so I try to replicate it on my character. Now, sometimes that references does not physically look like my character, but I have tried it anyway. Results have varied, but I have improved on making the result actually resemble my character even if the reference is very different. I would add that has worked for me if the character is humanoid. If the character is a monster, so I try to find references from monsters. Also it has applied, at least in my case, when I have designed mecha like characters, like gundam giant robots. The thing with mechas is the level of detail, they look in anime like the actual design of the action figure they will sell later, like if they were using the same very technical sheets. Everything ends up being an interesting, messy trial and error.
I've found that by lowering the Chaos level, it makes things look more normal or natural instead of some weird monstrous things
What else did you expect? XD
@@octavio675 Mentioned it because I didn't see the video acknowledge the chaos slider at all
@@valkyriagreyholme8135 Oh, well, I didn't watch the whole video so I don't even know what was shown and what wasn't xd
@@valkyriagreyholme8135 8:00
"Halfway Hitler, Halfway Jesus" Not a phrase i expected i would hear at any point today
9:29 It's changing shapes because the AI hasn't been trained to have memory of specific details, so it has to redo the entire image based on the slider values and original photos it has been given. So any change is a complete redo, even though it may appear similar. This is very common for current AI implementations because memory can have all sorts of unintended side effects.
I might use it to start sketches from the generated characters or make tokens for roll20 games, so thanks for covering it Boro :)
How some people actually think digital art is made ,🤔
Never heard that
I’m confused. What do you mean?
I have met plenty of boomers who think that CGI and digital art isn’t art and isn’t result of hard work and time. They think someone just pushes a button and the computer does all the work.
@@MildMisanthropeMaybeMassive not to say that making this technology isn't an art form in itself, but yes i agree. But i think this was a joke, and some can relate and others can't relate that much. So i just hope people don't take it to serious 🤣. And yes i should probably have written" how some" Instead of generalizing.
@@epicepic4561 Well, Art is like the most subjective shit there is so there that
There's so much you can do with that it's almost scary
Sounds fun to me
@@sauviel6296 Yeah it's fun until most of us artist become obsolete and lose our jobs.
@@Walperion_Music not really, you still need to know what looks good and what doesnt, what works in videogames, how to adapt this images to every proyect and so on. Concept artist have used any tool they could to come up with more ideas and iteration, You can watch the 2013 !! video "inspiration and techniques " from scott robertson.
This just help artists to be more creatives and faster. But it doesn't show's intent, colour theory, storytelling, function, etc. You still need artists for that.
@@Walperion_Music that is my fear
@@Walperion_Music Artists will never be obsolete as long as humanity is still around. Humans are adaptable. You never saw painters go extinct with the invention of the camera. You never saw traditional artists go extinct with the invention of digital art.
I use artbreed for quite a while now, whenever I'm writing a story I create characters and how should they look, so it helps me to imagine their features better
That's awesome! Do you know if there is a way to put a portrait face into the full body creation?
@@martfp88 i think the closest you could get is to photoshop a portrait to a full body
This is so insanely amazing. And the terms make it Creative Commons and public domain, so you can literally use this for anything you like; CD covers, board or card games, artwork, etc. I am SO using this for some of my creations. Thanks for the video on this.
I’ve used this a ton actually since I can’t afford commissions for actual art in order to at least get concept creations for a world build. It’s been awesome.
this site is amazing wth, it's like an unlimited source of face references
I recently started toying around with it and frankly I am impressed. Not only was it pretty entertaining, but I found it as a great source for finding a foundation for a concept, either as a reference or as something to paint on top of.
"Ok, I give up, I can't make Jesus in here" DANTE, BORO.
Alfred E Newman of Nazareth. "What Me, Worry?"
*proceeds to make Jesus-Hilter*
@@user-kb7kf4vu2u I already did that
For artists, this seems like a really good tool for iterating ideas. For fakers, just one more way to fake it.
i saw one account literry post these on their instragram account for years and i used to follow then but they dont exist anymore
@@hatpal2287 Edmond Amario?
@@philkinney9015 pretty sure
@@philkinney9015 i just found out he changed his username to art zuru
I used it to see what my characters would look like if they were real people. It was….interesting to say the least haha
"Is there a genital slider?"
That's a furry art, you don't wanna have a genital slider.
no that's why he asked
Yes, but ... What was the answer?
honestly this video made me feel a lot better that I couldn't figure out how to make portraits look good on ArtBreeder lmao I thought I must be stupid or missing something and kept trying to find tutorials... months later the algorithm gives me this video and I'm happy to find that you had exactly the same experience I had haha
I'm glad you mentioned the website's name. Cause literally every is like... Ah I made this with Ai... Like wtf Ai? Just tell me where you did that? 😂😂😂
They’re trying to sound smart or something lmao
@@aleph4425 Exactly 🤣 just say I made this by moving a few sliders around.
FYI, this appears to be an evolution of ganbreeder, which is available on GitHub. Whenever possible, run ML locally, because 99% of these projects seek to monetize your input. It doesn't look like ganbreeder has had updates in awhile so YMMV. Hope this helps.
Other meaning of Ai is adobe illustrator
@@sorashiro2359lol, I would know because I use it every day. XD
Now this is going to become an unlimited fountain of D&D npcs for me.
Holy crap this is a strong tool! I can see a lot of artist will start to use this for ideas and to get faster to the end result.
I have no art skills or anything, just randomly stumbled in here and this was extremely funny to watch
I've been using it in projects to quickstart backgrounds and stuff for pics. The algorithm is sort of limited, but still extremely useful
"--There's too much happiness for the human flesh so it starts dissolving."
Happens to me all the time.
for an artist it's like a land of hope for a new inspiration
I'd say the results become _more_ interesting when they start falling apart.
You sound like my thoughts when I tried this out a while ago. "Hmm maybe a bit less, nope that's too much human, why tf do you change the shapes?!"
Putting your own images trains the algorithm with a subscription. It clearly says it will not represent your art that you upload but rather it makes an interpretation to use.
If you have a paid private account, and upload enough of your own art, then tag it correctly, you start to see the results when it makes more art like yours.
This is pretty insane. I don't feel comfortable with a future where no skills are required anymore to create art. I feel like an old men yelling at clouds
Heh cloud
AI is only able to imitate and combine art that is drawn by a person, it is not able to create something new. In any case, we will need those who will feed the AI new styles and new character designs, etc.
@@orgax Eventually it will be able to create something new. I would argue it is already doing that here. Human artists are also inspired by other artists, styles, influences etc. Nothing that couldn't be simulated as well at some point.
@@rickysargulesh1053 This requires a conscious AI, which is almost impossible.
@@rickysargulesh1053 well it still needs a human to put it to work and give meaning to that work. But if some day not even that is needed i think the biggest problem will be that the machine might have be alive rather than human artists not being as unic anymore..
I've used this website before, I found it quicker to create and mix faces (I tried portraits only) with the "Parents" option. You select up to 8 images I think and mix them and you can add different characteristics with the sliders as well.
thanks for sharing this cool web site. I'm quite fond of AI when it goes crazy. I'll have a lot of fun exploring this.
This looks great I would absolutely use this to come up with ideas for 3D characters to sculpt.
"maximum human is too much human" was my favorite part
12:19 this is indeed the first video I've watched from your channel, so glad to know what I'm getting into.
"there's too much happiness for the human flesh" I'm stealing this quote for my game
if I remember correctly, artbreeder has a thing where you can use whatever image you made for whatever you want. so like your image is basically yours, and you can use it as character design or album covers! but i could be remembering wrong
Me : *Just drawing n painting*
AI : I'm gonna end this man's whole career
Have been looking for an app like this. I am a big fan of AI assisted artwork. Without really searching for it I had this video in my recommendations. This is really helpful. Will give this app a try, looks exciting to play with. Thank you for making this video!
I love how clean your audio sounds, great work with your acoustic treatment
The Album cover generator on that website is the best part
Props for remembering Alchemy.
Character design is great. Issues though they should add before commit to purchase:
Great find, this online app is awesome!
Posing - alternate positions
Saving building block - ability to recreate the original AI version, currently it seem too random when starting over or there is no option to save settings or revert to last setting (ie history or save points)
at other things too which will go into later.
Those AI face generator are too good
I love your weird and subtle sense of humor lol, cracks me up
Chaos is basically how much impact do you want your sliders to affect the base picture. No chaos gives you the base picture, high chaos makes the photo very sensitive to your slider changes.
This is impressive. Probably a cool way to generate new faces for references
Oh this is pretty awesome! thanks for the info, definitely going to be using it for crazy expressions. Also, I finally subscribed cause of that Do-rag
ref 6:30 comment on happiness and darker tone. Over the past year I have been collecting images for the term "happy eyes". A few things struck me. 1. Kids genuinely let their smiles light their eyes; when you see it in an adult it's rare. 2.The most common pictures I've found of adults were "ethnitic" mountain people and herders who were untouched by the ways of man. 3.White people like to be posed and/ or stoic.( I literally had to tweak the search engines for a couple of hours to get candid shots. Happy little accident, it lead to discovering Henery Cavil who's eye still have their spark at almost 40.)
Also, great video keep it up. You're doing awesome!
Content means the pose and shape. Style means the color pallet and the style it is drawn (ie style of lineart, how the coloring is done, etc).
Played with it for a few minutes. Mixing a landscape with some sci-fi texture can get some cool stuff really fast.
In my experience with the sliders on Artbreeder you have to lean very slightly into them. Just going .1,.2,.3..etc I don't usually go beyond 1 or 2. It goes too extreme into the categories when you go crazy with the sliders.
I tried out this tool after watching your video. I'm definitely liking how it helps with idea generation!
Aside from the cool concept, you're funny as hell. New sub.
I love the bandana Boro lol
In case anyone has been wondering why those sliders have such weird effects: The Designer of that website did not decide which slider is supposed to do what.
Basically, the AI was given a set of images and it figured out how many distinct traits there are between the images. They just named the sliders afterwards, depending on what they did, hence why one slider can affect another trait also.
carykh built a similar system, merging portraits of celebrities and explaining what's going on, in case anyone wants to know a bit more.
Thanks! Lots of fun, and maybe some serious work if, um. worked at. As someone who likes imagining alien landscapes, and critters, I'm going to to some determined work with this. Thanks for the upload and intro!
If your creatively void like me then this can be a nice tool to get started for concept art, for say a game or whatever.
dayum those character design was badass wish i can do that details
thanks for sending me for a generative spiral for the whole evening!
There were so many golden moments in this video
This is mind blowing. Soon computers will be making whole videogames by themselves at a rate of millions per minute and we just will have to vote for the good ones.
Dude this literally was what I was asking for today. You just solved so many of my problems
Now I cant unsee whatever I saw on the site, Thanks
You should try Ebsynth. It lets you animate paintings by merging it with video.
this video was recommended to me 5 times, i have 0 regrets
The landscape one is gonna help me a lot. I kinda suck at designing landscape.
This reminded me of Orwell's 1984, when he talks about how art an entertainment products are made automatically
Well...nuff said
Games in the future where the quests are build around what you do by a.i, would make any game pretty much endless. gta where you can really do what you wanna do.
@@Monkeyking112 we already have the prototype of that in the form of AI dungeon
⠀⠀⠀⠀⠀⠀⠀⣠⡀⠀⠀⠀⠀⠀⠀⠀⠀⢰⠤⠤⣄⣀⡀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⢀⣾⣟⠳⢦⡀⠀⠀⠀⠀⠀⠀⢸⠀⠀⠀⠀⠉⠉⠉⠉⠉⠒⣲⡄ ⠀⠀⠀⠀⠀⣿⣿⣿⡇⡇⡱⠲⢤⣀⠀⠀⠀⢸⠀⠀⠀1984⠀⣠⠴⠊⢹⠁ ⠀⠀⠀⠀⠀⠘⢻⠓⠀⠉⣥⣀⣠⠞⠀⠀⠀⢸⠀⠀⠀⠀⢀⡴⠋⠀⠀⠀⢸⠀ ⠀⠀⠀⠀⢀⣀⡾⣄⠀⠀⢳⠀⠀⠀⠀⠀⠀⢸⢠⡄⢀⡴⠁⠀⠀⠀⠀⠀⡞⠀ ⠀⠀⠀⣠⢎⡉⢦⡀⠀⠀⡸⠀⠀⠀⠀⠀⢀⡼⣣⠧⡼⠀⠀⠀⠀⠀⠀⢠⠇⠀ ⠀⢀⡔⠁⠀⠙⠢⢭⣢⡚⢣⠀⠀⠀⠀⠀⢀⣇⠁⢸⠁⠀⠀⠀⠀⠀⠀⢸⠀⠀ ⠀⡞⠀⠀⠀⠀⠀⠀⠈⢫⡉⠀⠀⠀⠀⢠⢮⠈⡦⠋⠀⠀⠀⠀⠀⠀⠀⣸⠀⠀ ⢀⠇⠀⠀⠀⠀⠀⠀⠀⠀⠙⢦⡀⣀⡴⠃⠀⡷⡇⢀⡴⠋⠉⠉⠙⠓⠒⠃⠀⠀ ⢸⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⠁⠀⠀⡼⠀⣷⠋⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⡞⠀⠀⠀⠀⠀⠀⠀⣄⠀⠀⠀⠀⠀⠀⡰⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⢧⠀⠀⠀⠀⠀⠀⠀⠈⠣⣀⠀⠀⡰⠋⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
So assuming they use generative adversarial networks, the way they make these sliders is probably by having people label the training image datasets, so they have a dataset of thousands of the input original art where someone has to decide what value in each category to assign to it, e.q. 0.5 in "style" or whatever. Then the model will try and learn to find the pattern and what changes based on the "style" variable. I would imagine that this is why the open eyes variable works so well, mostly because it's fairly easy to unambiguously assign a value of of how open the eyes are, e.q. 0 at completely closed and 1 at wide open or something, and this will make it straightforward for the model to make sense of, whereas some of the more ambiguous variables are tougher to judge and will depend on how the people labelling them judged them to be valued, which will introduce some contradictions and subjectivity, making it easier for the model to get confused.
I think it's more likely that they've postfactum manually selected vector components/sets of components that worked for changing human-meaningful features
@@dodecahedron1382 Maybe, I don't actually have practical experience with implementing GANs, but from my understanding, you'd have to later poll a countless number of images in the output image space and then average the corresponding vectors just to calculate the average input image feature vectors, no? You'd have to have supervision either way, and it sounds to me like the former way is significantly less messy. But as I said, I don't actually know, so maybe you're right.
That's really interesting. Thank you.
A sophisticated version of the old ‘Morphing’ programmes.
Looks like what i generate in my head at times. Like I can think of what i want but can't physically draw or explain what im seeing or imagining. This could be great for ideas when creating a character. You could basically make your weirdly generated thoughts and ideas to the screen (in an ugly deformed way) and then refine those details now that you have your thoughts and ideas physically in front of you for reference.
Ok maybe it doesn't generate your actual thoughts as its the AI doing the thinking. But just imagine if there is a generator that creates images from your mind?
Give it 10 years at least. Then it's possible
"Too much happiness for the human face, so it starts dissolving." LMAO! Best sentence in this video.
Second best comment, "Maximum human is too much human." :-D
Third best comment, "Welcome to BoroCG channel where we crossbreed furries." Hahaha!
Haha yes, I also think these are the best.
I used this to make the charaters from my books using sketches I made of them. It's really cool what it came up with. (I had to do some edits to it after I was done. But still amazing as a helping hand.)
Looks pretty cool
gonna have to try it for myself but it looks like it would be great for making reference pictures
Thanks.. before this video i was really bored... now i had an "interesting" evening finding out a lot about my type of aesthetics!
I've actually used artbreeder for so many creepypasta type pics, it is soooo amazing!!!!
"too much fun for the human flesh" - ROTFL. Bro, you are funkin hilarious. I loved this video, its the nexus of art, tech, irreverence and chill, and had me cracking up. Thanks.
This worked better than expected, its a great website for making your own references without having to worry about other people getting mad at you for using someone elses work
The thing is, based on how Neural networks work, each slider is responsible for a value somewhere in the network brain. Now, the problem is, that those values are not only not labeled and interconnected, but also don't make any human sense. Trying to change them is like trying to affect human brain by poking neurons one by one.
So, usually developers just try changing them and then take those that have visible effect to make them into sliders.
Second thing is, problem with uploading photo lies in the fact, that nns need to convert picture into a set of features that they can understand, that they figured out while being trained. If photo contains features that weren't there while trained, they are skipped or even break something.
There needs to be a feature where it creates a character turn around for you and you can make small adjustments as needed in a different program. That would save me so much time
That would be awesome. You can do a small amount of that for the front view by using the crossbreed feature - just crossbreed a character facing the opposite way, crank the Style slider all the way toward your original character, and then use the Content slider to control the pose. Kinda tricky but at least it can give you a small amount of different angles for the front views.
@@DonnieMaynardChristianson Thank you so much man I'll definitley give that a try :)
Hey there! I thank you for showing me Artbreeder. It is a good tool to work on templates/WIPs for actual work later. Greets from Germany, Holger.
I hope you do a follow up where you take some of these designs and do a full piece with them
This is amazing Site, A Big Thank You for this information .
7:00 "I want you to do it again, but I want you to do it happier and with your mouth open."
I learned stuff by just listening how you describe the changes, oily skin, flash photo, studio light.
pretty nice nightmare fuel at high numbers
Fun tool. Very good for inspiration.
6:18 "interesting how thats connected to happiness" made me laugh out loud
"And that's where you get some horror stuff" XD
After reading Doraemon as a kid i always think "man, wouldn't it be great if there's a robot that can draw my ideas?" Aaaand there is now.
i think this is the perfect tool for creating museum over museum filled with questionable modern art
*AI - “All your colored pencils belong to us”*
Concept Artists: They took our jobs!! 😨
I read "Field Day" all over this video and app. The Ai as usual has much to learn, but this is fun to mess around with. I remember using another app like this and now It's time to have a bit of fun.
this is crazy magic!
Wow, now I'm terrified for my art career in the future.
you can actually get the sliders pass 2 if you hover over the above the slider theres a up and down arrow that passes the maximum