Best Apps For Stunning AI Art (Midjourney Alternatives)

Summary

The video provides a comprehensive ranking of popular AI art generators, with Midjourney V5 rated as the top choice due to its superior aesthetic quality and handling of human features. Alternatives like DALL-E 2, NightCafe, and Leonardo.ai are discussed for their unique strengths and free usage options. The video also covers other tools like Stability AI, Canva, and Runway, highlighting their specific use cases. The speaker explains how to make money with AI-generated art through platforms like Wirestock, and encourages participation in creative AI art events hosted on Discord.

Takeaways

• 🎨 Midjourney V5 is considered the best AI art generator, especially for photorealistic images, with significant improvements in generating human features like hands, teeth, and faces.
• 💰 Midjourney is now a paid service, and the free trial has been removed, though sometimes it is briefly re-enabled.
• 🖼️ DALL-E 2 by OpenAI is an A-tier AI art generator, particularly strong with abstract and cartoonish art, but not as effective at generating realistic human characters.
• 🧑‍🎨 NightCafe is user-friendly and offers good quality AI art, but struggles with generating human figures, especially complex body postures like ballerinas.
• 🎮 Leonardo.ai is a highly recommended free-ish AI art generator, especially good for generating humans, with access to multiple models and the ability to fine-tune them.
• 💻 Wirestock allows creators to sell AI-generated art on major stock marketplaces like Adobe Stock, Pond5, and DepositPhotos, taking an 85% cut for creators.
• 📈 Beginners can use platforms like Dreamline Art and Deep Dream Generator, which offer artsy and user-friendly options but may not match the quality of higher-tier services like Midjourney.
• 💡 Stability AI’s Dream Studio allows fun features like auto-expanding images, but like most Stable Diffusion models, struggles with human characters.
• 📷 Canva’s AI image generator, despite being convenient, produces poor results when generating human images, earning it a C-tier rank.
• 🎥 Runway is notable for its text-to-video features, but its text-to-image results are below average, particularly with generating human characters.

Q & A

What makes Midjourney V5.1 the top AI art generator according to the script?

Midjourney V5.1 is considered the best due to its ability to generate highly aesthetic, photorealistic images, particularly when it comes to human anatomy, faces, hands, and other challenging details. It stands out among competitors for its superior quality.

Why did Midjourney remove its free trial, and how does this affect users?

Midjourney removed its free trial, which previously allowed users to generate 25 images for free. This change forces new users to pay for the service. Occasionally, the free trial is re-enabled temporarily, but it is typically a paid service.

What differentiates Dall-e-2 from other AI art generators?

Dall-e-2, created by OpenAI, uses a custom model rather than stable diffusion, which is open source. This unique model gives Dall-e-2 its own capabilities, particularly excelling at cartoonish and abstract images, but it struggles with generating photorealistic humans.

How does Night Cafe rank in terms of user experience and image quality?

Night Cafe is praised for being very user-friendly, with presets, competitions, and examples to inspire users. While it excels at generating certain types of images, it struggles with photorealistic humans, especially in complex compositions like body anatomy.

What are the strengths of Leonardo.ai, and how can users access it?

Leonardo.ai offers 150 free credits daily and excels in generating human-like images. It allows users to fine-tune models and has a wide selection of styles, including a diffusion model with dark and aesthetic qualities. Users need to sign up for early access, which takes 3-5 days.

Why is Stable Diffusion often criticized for generating human images?

Stable Diffusion, which underpins many AI art generators, tends to struggle with rendering realistic human anatomy, especially in complex poses or faces. This often leads to distorted or unusable images when trying to depict people.

What features make Wirestock an attractive option for selling AI-generated art?

Wirestock simplifies the process of selling AI-generated art by uploading images to multiple stock marketplaces simultaneously, like Adobe, Pond5, and DepositPhotos. Creators get an 85% cut of sales, and Wirestock is developing an AI upscaler to help artists meet resolution requirements.

Why does the video place Canva's image generator in the C tier?

Canva's image generator is based on Stable Diffusion, which struggles with human images. However, it's ranked in the C tier due to its convenience, being integrated into Canva's design tools, which makes it useful for simple tasks.

What does the script suggest about Dream Studio by Stability AI?

Dream Studio, created by Stability AI (the company behind Stable Diffusion), offers fun features like auto-expanding images. However, the script suggests that while it can be enjoyable for enthusiasts, it lacks the quality found in top-tier generators like Midjourney.

What are the downsides of lower-tier AI art generators like Crayon or Lensa AI?

Lower-tier generators like Crayon (based on Dall-e Mini) and Lensa AI are limited in quality. Crayon delivers simplistic and often underwhelming images, while Lensa AI, which became viral for its photo remixes, is inferior to Midjourney in terms of quality and flexibility.

Outlines

00:00

🎨 The Rise of AI Art Generators and Midjourney's Dominance

AI art generators have surged in popularity due to their utility and entertainment value, but not all tools are created equal. The best current option is Midjourney V5, which excels at generating aesthetic and photorealistic images, particularly with human hands, faces, and bodies. However, Midjourney has recently removed its free trial, which may limit its accessibility. This paragraph introduces the concept of AI art generators, the tier list ranking system, and establishes Midjourney's dominance in the field.

05:03

💰 Midjourney’s Cost and Free Alternatives: Dall-e-2, Night Cafe, and Leonardo.ai

Midjourney, while the best, is no longer free, prompting many users to look for alternatives. Three notable free options in the A-tier are Dall-e-2 by OpenAI, Night Cafe, and Leonardo.ai. Dall-e-2 is based on a custom model different from the widely used Stable Diffusion. It's great for abstract and cartoonish images but struggles with human realism. Night Cafe, known for its user-friendliness and social features, is easy to use but has issues generating realistic humans. Leonardo.ai, although marketed for game asset generation, offers strong results, especially for human images, thanks to its customizable models.

10:04

🖼️ Leonardo.ai and Making Money with AI Art

Leonardo.ai stands out due to its customizable models and daily free credits. It produces good results, especially for human figures, and allows users to fine-tune models with their own data, making it versatile for many types of art. Additionally, this section introduces Wirestock, a platform that allows creators to sell their AI-generated art on multiple marketplaces. Wirestock simplifies the process of selling digital art, offering up to 85% of the profits, and will soon introduce its own AI upscaling tool to ensure image quality meets marketplace standards.

🧑‍🎨 B-Tier AI Generators and Niche Use Cases

The B-tier includes AI art tools that are useful for specific cases but lack the versatility of higher-ranked tools. Dreamlike Art offers a user-friendly interface for beginners, while Deep Dream Generator produces artistic results but has a high subscription fee. Stability AI, which developed the widely-used Stable Diffusion model, provides Dream Studio, a free platform for experimenting with auto-expanding images. These apps are fun to use and great for enthusiasts, but for premium performance, Midjourney and Leonardo remain superior.

📉 C-Tier and Lower AI Tools: Usability Issues and Drawbacks

C-tier tools like Canva’s AI generator and Crayon offer convenience but fall short in performance, particularly with human figures. Canva integrates Stable Diffusion but is mainly useful within the context of its design suite. Below the C-tier, apps like Lensai and Jasper are either outdated or underperform, especially compared to free alternatives in higher tiers. These apps may appeal to niche users, but their value is limited in comparison to top-tier tools.

🎉 Fun with AI Art Competitions and Final Thoughts

The video closes with an invitation to join the creator's Discord server for AI art remix competitions, showcasing the creative possibilities of AI art generation. Examples of winning entries, including a SpongeBob remix, are highlighted. The conclusion also encourages viewers to leave feedback on the AI art tool rankings, which will be regularly updated as tools evolve. The final thought emphasizes Midjourney as the most reliable paid tool, with Leonardo as the most effective free alternative.

Keywords

💡AI Art Generators

AI art generators are software or platforms that use artificial intelligence to create images based on textual descriptions or other inputs. The video discusses various AI art generators and ranks them based on their quality, usability, and cost. Examples include Midjourney, DALL-E 2, and Night Cafe, which offer different styles and functionalities for generating art.

💡Midjourney

Midjourney is highlighted as the top-tier AI art generator (S tier) for its ability to produce photorealistic and aesthetically pleasing images. It has its own proprietary model, which performs exceptionally well with human anatomy, facial details, and hands—common challenges in AI art generation. However, the video mentions that Midjourney recently removed its free trial, making it a paid service.

💡DALL-E 2

DALL-E 2 is an AI art generator developed by OpenAI. The video categorizes it in the A tier, explaining that it is highly effective for abstract and stylized art but struggles with generating realistic human characters. DALL-E 2 differs from other models like Stable Diffusion because it is based on a unique architecture developed by OpenAI, making it distinct in the field of AI art.

💡Stable Diffusion

Stable Diffusion is a widely-used, open-source AI model that serves as the foundation for many AI art generators, including Night Cafe and Leonardo.ai. While it is accessible and versatile, the video points out that it often produces poor results when it comes to human figures, due to difficulties in accurately rendering facial features and body proportions.

💡Leonardo.ai

Leonardo.ai is a versatile AI art platform that offers 150 free credits daily to users who sign up for early access. The video places Leonardo in the A tier for its usability and quality of results, especially when generating human images. It allows users to fine-tune their own models, making it a powerful tool for creators looking to customize their output.

💡Night Cafe

Night Cafe is another AI art generator that uses the Stable Diffusion model. The video ranks it in the A tier due to its user-friendly interface, daily challenges, and community features. Despite its ease of use, it is less effective at rendering realistic human figures compared to other generators, which limits its versatility in some use cases.

💡Tier List

The tier list is a ranking system used in the video to categorize various AI art generators based on their overall performance, usability, and cost. The tiers range from S (best) to F (worst), with Midjourney placed in the S tier, DALL-E 2, Night Cafe, and Leonardo in A tier, and other less effective tools in lower tiers. This format helps viewers quickly identify the strengths and weaknesses of each tool.

💡Human Figure Generation

Human figure generation refers to the AI's ability to create realistic images of people, including facial features, anatomy, and body posture. The video emphasizes that most AI art generators struggle with this aspect, often producing distorted or unusable results. Midjourney is praised for excelling in this area, while others like Stable Diffusion are noted to perform poorly.

💡Wirestock

Wirestock is a platform mentioned in the video as a sponsor. It allows creators to upload and sell their AI-generated art on multiple stock photo sites like Adobe Stock and Pond5 without the hassle of multiple sign-ups. The video explains how Wirestock simplifies the process of selling AI-generated content by handling distribution and offering an 85% revenue cut to creators.

💡Custom Models

Custom models refer to AI models that are uniquely trained or fine-tuned by companies to provide distinct outputs, different from the standard versions like Stable Diffusion. Midjourney and DALL-E 2 are highlighted as examples of platforms using custom models, which allows them to produce higher-quality and more specialized results compared to generators based on generic models.

💡Upscaling

Upscaling in the context of AI art refers to increasing the resolution of an image while maintaining or enhancing its quality. The video mentions the need for upscaling AI-generated art before uploading it to platforms like Wirestock, as many stock photo sites require high-resolution images. Upcoming tools for automatic upscaling, such as those integrated into Discord, are also discussed.

Highlights

AI art generators have become increasingly popular and offer useful, fun tools for generating images.

Midjourney is widely regarded as the best AI art generator due to its aesthetic quality, ability to handle human anatomy, and photorealism.

Midjourney no longer offers a free trial, making it a paid service.

Dall-e-2 by OpenAI is another popular AI art generator known for its high quality, particularly with abstract and cartoonish images, but struggles with human figures.

Night Cafe is highlighted for its ease of use, offering daily challenges and a social feature that makes the platform user-friendly.

Night Cafe and Stable Diffusion models struggle with generating realistic human images, often producing unsatisfactory results.

Leonardo.ai is noted for being free-ish, offering 150 credits a day, and provides good results for human images by allowing users to pick from various models.

Leonardo.ai also allows users to fine-tune their own models, making it a versatile platform for creators.

Wirestock, a platform for selling AI-generated art, offers a simple way to distribute images to major stock marketplaces.

Wirestock's AI upscaler will soon make it easier to upload high-resolution AI art, integrated with Discord for seamless image sharing.

The video highlights lesser-known AI art generators like Dreamline Art and Deep Dream Generator, which are easy for beginners to use.

Stability AI’s Dream Studio is fun to use for features like auto-expanding images but isn't as powerful as Midjourney.

Runway is noted for its exciting work in text-to-video, although its text-to-image capabilities are average.

Crayon is based on the Dall-e Mini model and generates lightweight, less detailed images.

A Discord community hosts AI art remix contests, fostering a space for creators to share and compete in AI art generation.

Transcripts

AI Art generators have been all over the  Internet lately and for good reason. They  

can be really useful and they're actually a lot  of fun. But not all of them are created equal  

and this can make it really hard especially  now that Midjourney removed their free trial.  

So if you ever wondered what other alternatives  are actually out there and which ones of those  

are actually useful, you're in the right  place because now I'm going to show you  

a ranking of my favorite AI art generators  and a bunch of examples to help you decide.

So let's get started from the top by looking  at the very best ones first. And as you might  

already know, the best one right now is Midjourney  V5 and now 5.1 is even better so that's why it's  

S tier. And if you're not familiar with these  tier lists, it's very simple. This is similar  

to grading with school with S tier being just  perfect. So down here F is the worst and S tier  

is the best. And the reason Midjourney is the best  is because it's very aesthetic and they manage to  

fix the issue with human hands. They mostly fix  teeth, bodies, and human faces and they're also  

the best at generating photorealistic images.  And I don't think this is a large discussion,  

most people that use AI art extensively would  agree that Midjourney is King right now.

Alright, so Midjourney is best but what's the  problem? Well, it's quite simple. They charge  

for their services and they now remove their  free trial so you don't get 25 images for free  

anymore as it did over the course of the last  few months. If you follow the server closely,  

sometimes they re-enable it for a few  days but usually, it's going to be paid.

And that brings us to our next tier here because  most of these underneath are actually free to use.  

And in our A tier category here we have Dali 2  by OpenAI, we have Night Cafe Art Studio, and we  

have leonardo.ai. And all three of these are very  different so let's talk through them one by one.

So Dall-e-2 is a model created by OpenAI  and it's particularly interesting because  

Dall-e-2 is actually based upon a custom model.  Most of these other ones are based upon stable  

diffusion which is the most popular AI image  generator that is fully open source so anybody  

can build on it for free. Dali and now Dall-e-2  is different because they build their model. And  

if you didn't know also Midjourney is different  in that way and that's why they're better. Any  

further details are not known to the public  but all their status they have their own model  

now which seems to be better than all the stuff  out there and that's why they're doing so well.

So how do you use this? Well, you're gonna need  to open an AI account just like you do with  

ChatGPT and then you can go to labs.openai.com  and start generating images with Dall-e-2. And  

there was a waitlist for this first but around  the end of 2022 they just put it out to the  

public and everybody can use it. So you can  upload images, you can take inspiration here,  

or you can just use the text image features  in this text box and that is pretty much Dali.

And the quality here is very high especially when  it comes to cartoonish or abstract items. But I do  

have to say this when it comes to generating  human characters like anything that involves  

humans within the picture which is actually one  of the most common motifs that you will want to  

generate in a lot of the cases, Dali 2 is not  great. And what we did is we generate the first  

two example prompts from our resource here. So  look, this right here, the results you would get  

with Midjourney and here is what you get with  Dali 2. It's already no detail in the phase and  

if you look at the ballerina this is just very  different. So if you're looking for the stylized  

look within humans it's fine but if you're  looking for photorealistic images of humans,  

none of the models presented here are as good  as Midjourney right now. That's just a fact.

But with that being said, let's move on and we'll  keep these two examples in mind as we re-run them  

for all of these image generators. Okay, so  next one on the list is Night Cafe and this  

one is actually super easy to use and yields  very good results. All you need to do is go to  

nightcafe.studio and again all the links will be  in the description below and all you need to do  

is with Night Cafe is just create an account  and you can start generating. And even these  

daily challenges are kind of fun to participate in  and as you can see these results are really good.

So while crafting this tier list we  also considered user friendliness and  

Night Cafe is very user-friendly because it  offers some of these presets, competitions,  

and a lot of quality examples for you to get  inspired. And what intrigues me personally  

about this is that they have this social  feature plus the quality of this image in  

particular is just excellent. This is a 10  out of 10 when it comes to generating this  

type of image and that's what I love here.  It might not be able to do everything but  

the things that it can do are excellent  and that's why I would recommend it.

And for Night Cafe, if we do our little comparison  here, it's just gonna be absolute catastrophe for  

the app because it's based upon stable diffusion  and stable diffusion by default is terrible at  

humans. It will create these monstrosities that I  will show you now so just be prepared. So here's  

what it did for the image of the homeless  woman. As you can see the environment is  

okay and actually quite good but when it comes to  showing the people it's absolutely unusable. It's  

at least smart enough to cover up that fact  in some of these other images. And when it  

comes to ballerina that is kind of the ultimate  test because you have to get the anatomy right,  

you'd have to get the face right and usually  the body postures are quite extreme but they  

need to make sense otherwise us humans will  pick up on the fact that it's not right. So now  

have a look at the atrocity that it produced.  Well, there's like 50 things to critique here  

so we'll just close it off and now you know  that this is not meant to be used for humans.

Okay, let's move on to the next one. So the next  one is leonardo.ai. And Leonardo is interesting  

because it's free-ish currently. It comes with  150 credits a day if you get an account and in  

order for you to get an account, you need to  sign up to Early Access with your email. From  

what I've heard usually takes three to five  days for you to get accepted and then you  

can start using your free accounts every day.  And don't get thrown off by the fact that they  

market this to be a game asset generator. This  is excellent for so many other use cases too.

So again, they're really marketing this towards  game designers but if you look at the examples  

of the homeless woman and the ballerina you  will find that Leonardo yields some of the  

best results here. And the reason for that is  that once you're inside the app, you get to  

pick from various models and some of them are  actually quite good with humans. Look at this,  

this is not perfect but it's quite good and I  would say the homeless woman is actually quite  

usable. So Leonardo is really great and what  you'll see once you're inside is that you get  

all these various models to pick from. So Leonardo  Diffusion is a fine-tuned model of table diffusion  

that is dark and very aesthetic in my opinion.  I really like this one and with this deliberate  

model, you will actually be able to create  decent humans and there are many more in here.

One more thing that I should point out is  that it actually allows you to fine-tune  

your own models. So I could actually take  one of these models and train it further  

with my own data which is absolutely amazing  and it's so user-friendly on here. Honestly,  

this is shaping up to be one of the best AI  art platforms as their mission is trying to  

be the Photoshop of AI art. So I strongly  recommend you check this one out and again  

and get a bunch of these credits which  is more than enough to get you started.

And now I want to take a moment to tell you  about a possibility that you could use to  

make some extra money on the side with your  AI art generations and that is for today's  

sponsor wire stock. It's the easiest way to  sell your AI-generated art with over 230,000  

creators and more than 9 million media  items. Wire stock makes it a breeze to  

upload and sell content on the world's largest  stock marketplaces. And as we are discussing  

AI-generated images today it's important  to mention that not all of these platforms  

except AI generated alright but with wire  stock, you get to upload to Adobe Photo  

pond5 and deposit photos simultaneously.  All of those three are good to go and if  

you sell images on there you're gonna get an  85% cut. Additionally, creators get individual  

portfolios for free where they can sell content  directly, create collections and even sell prints.

But it should be noted that you need to upscale  your images before you upload them to wire stock  

otherwise they won't be accepted. And right now  there's many tools to do that but very soon there  

will be releasing their very own AI upscaler  that will help you upload to their platform.  

That's going to make it super easy to enhance  and upscale your AI art so it has the necessary  

resolution for all of these sites to accept it  and guess what they're going to be integrating  

that with Discord which will allow you to directly  send images from Discord to wire stock and they  

will distribute it to the part-time marketplaces  making everything as seamless as it could be.

So here's the deal with wire stock, you get  direct access to selling on the world's largest  

marketplaces without the hassle of separate  signups, extensive forms or payment on multiple  

sites and all that for a 15% fee. As someone who  has been uploading and selling stock photos and  

videos for seven years let me tell you that's a  good deal especially when generating some of these  

images takes seconds and then distributing them  amongst the platforms will take beginners hours.  

So if you're an AI artist or just looking to sell  your creative work online give wire stock a try.

Click the link in the description  to try it out for free and if you  

decide to upgrade make sure to use  my code advantage20 for a discount  

at checkout. And now let's get back to  exploring the best AI image generators.

The next category, the B tier, are ones that  I have bookmarked but I only come back to  

them for specific use cases. So for example,  dreamline art is a very beginner-friendly way  

to get started with all of this and it gets  you up and running with stable diffusion like  

no other app extremely beginner friendly  and I would strongly recommend this one.  

Deep Dream Generator on the other hand is  a fantastic app, with very artsy results,  

but their plans start at 19 a month and for  that, I would just go straight to mid-journey.

And then in this tier I also put Stability AI  which is actually the company that brought out  

the stable diffusion model that I talked  about so much and that most of these are  

actually built upon and they have their  very own app called Dream Studio which  

you can access for free and this allows you  to do some fun stuff like auto expand images  

and it's quite good at that and just fun to  play with. So if you're an AI art enthusiast  

I would recommend you play with all seven  of these and see what they're good at,  

what those different models are capable of and  where the weaknesses are. And if you're gonna  

pay for one I would make that Midjourney with the  most useful free one being Leonardo right now.

And that brings me to the last category. So here  we enter hmm that's kind of interesting territory  

but honestly I don't really return to many of  these because they have major weaknesses and in  

the categories above you will usually find better  results that are also all free. So for example,  

the Canva image generator yet again is based upon  stable diffusion and will get you results like  

this or this when it comes to humans so absolutely  unusable but it is inside of Canva and we actually  

use Canva to design some of our Instagram posts  and there it's just super useful to have it right  

there. So for the sake of convenience and ease  of use, it actually got C tier although it's  

really not that great. So the same thing would  go for most of these, they're either easy to  

use or integrated into something interesting and  this is the category where I would say if you're  

interested in this stuff check all of them out  because they could introduce one upgrade and all  

of a sudden they would be great but right now  they're kind of in this weird in-between spot.

One more honorable mention should go to  Runway. They're doing some fascinating  

stuff in text to video but their text image is  just average at best. And I'll give them this,  

they actually generated the worst ballerina out of  all the examples. Have a look at this one, oh God,  

I'm sorry this is gonna haunt your  dreams now. And then we also have  

Crayon which is based upon the Dali  Mini model which is a lightweight  

model of Dali that we talked about above  and this generates results like this.

And look below this, we enter the territory of  the apps that tried but honestly are not very  

useful. I mean okay, text to Pokemon might have  some niche use cases but otherwise all of these  

I would rather avoid. I mean you might recall  Lanzai that was the viral app where everybody  

created AI variations of their own photos but  it cost money and it does nothing but generate  

images and Midjourney right now can remix your  photos way better than Lensai was ever able to.  

And also the Jasper image generator costs money  and it's not even as good as the free stuff in A  

and B tier so that's why it's down here. It's not  horrendous but I don't see a reason to use it.

And then the category down here are just the  ones that you should stay away from. It's  

not even worth your time exploring them. But  hey look, I might have missed something here  

so leave a comment below if you disagree with  one of these rankings but make sure to leave a  

reason and then over time we can update this tier  list to reflect the current reality including all  

of your guys' feedback to make this resource as  good as possible. As we'll make this amongst many  

other maternal resources available for free on our  website from here on out so you'll find a link to  

that below. You can check it regularly and we'll  be updating the rankings as these apps evolve.

And lastly, I would like to bring your attention  to the events that we started running inside of  

our Discord server. It's completely free, the  events are free, they're just there for you  

to have fun and we actually just ran our AI art  remix contest to remix various images. So with 50  

participants and I gave them SpongeBob and they  had five minutes to generate a remix and look,  

this is what the winner of this round, Karate,  came up with. Isn't that just incredible? So  

these are some of the possibilities here and right  here you can see the four winners. So this one was  

Success Kit before this was Nyan Cat and this  was the Chat GPT logo. So if you're interested  

in participating in the next one we're holding it  on May 20th and there's gonna be regular contests  

after that. So feel free to join us, generate  some AI art and learn while you're doing it.

I hope this video was useful and if you  want to learn even more about Midjourney  

and how to get the most out of it, you  will love this video which introduces  

you to a bunch of tips and tricks on  how to do interesting stuff with AI.