New Midjourney Alternative but UNLIMITED and FREE! (Tutorial)
ฝัง
- เผยแพร่เมื่อ 31 พ.ค. 2024
- In this video, I give a tutorial on using a new AI image generation tool alternative that is better than Midjourney in many ways and completely free and unlimited. You can easily create AI generative art that is more accurate to the detailed prompts you provide. You no longer need to be a prompt engineering wizard to get what you want. You can even train your own models or use community-generated models! Create everything you want including hyper-realistic images, video game assets, consistent character generation, icon sets, stickers, and more!
Enjoy :)
Join My Newsletter for Regular AI Updates 👇🏼
www.matthewberman.com
Need AI Consulting? ✅
forwardfuture.ai/
Rent a GPU (MassedCompute) 🚀
bit.ly/matthew-berman-youtube
USE CODE "MatthewBerman" for 50% discount
My Links 🔗
👉🏻 Subscribe: / @matthew_berman
👉🏻 Twitter: / matthewberman
👉🏻 Discord: / discord
👉🏻 Patreon: / matthewberman
Media/Sponsorship Inquiries 📈
bit.ly/44TC45V
Leonardo.ai
Contents of this video --------------
0:00 - Intro
0:19 - Overview of UI
1:17 - Featured Models
1:38 - Using a Model
2:10 - AI Image Generation Settings
3:26 - Community Models
3:36 - Icon Set Example
4:10 - Fine-tuned Models
4:53 - Community Models Pt 2
5:40 - Custom (self-generated) Models
6:24 - Prompt Generation / Ideation
6:55 - AI Canvas (beta)
7:45 - Settings / Options
7:59 - Image Examples
8:14 - Outro
My Workstation Setup:
Apple MacBook Air M2 - amzn.to/3GQFexg
LG Ultrawide 5k Monitor - amzn.to/3XsnBuC
Logitech Litro Glow - amzn.to/3HkP1wX
Vivo Monitor Stand - amzn.to/3Xv0TlU
Logitech MX Master S2 - amzn.to/3kyghiH
Logitech Craft Wireless Keyboard - amzn.to/3QSsHhx
Logitech HD Video Camera - amzn.to/3XMFFQc
Blue Yeti Microphone - amzn.to/3XICOaP
Uplift Standing Desk - amzn.to/3XMFYKQ
Apple AirPods Max Headphones - amzn.to/3XOwYF1
Large Black Desk Pad - amzn.to/3YdNChz
====================================
Subscribe - www.youtube.com/@matthew_berm...
====================================
Some of the links above contain affiliate links, thank you! - วิทยาศาสตร์และเทคโนโลยี
Your videos have been a game-changer for me. I've learned so much from watching them, and my skills have improved as a result.
Great content man, also like how you explain things so simply, definitely a like and a subscriber, keep it up!
Thanks so much! Let me know what other topics you want to see.
Great video Mathew. Can't wait to try this out. Thanks for the info.
Thank you! I’m making a video about ai music right now, hope you come back to watch it.
This is an amazing resource, thanks for sharing it! I will definitely be checking this out to play around with and maybe create some great tokens and character art for my TTRPGs.
Thank you!
:3 MidJourney they say is $10 minimum a month after 25 images. I've also heard it's free as long as you use the public threads. While MidJourney is cool to see others generate art as you do, it's tedious to scroll for and look for yours. It's just easier to have it appear like normal AI does.
I only use MidJourney for the URL link option for making similar art. Lol. Some other AI systems use URL and/or image to reference to, but as I've mentioned, they're not as good as MidJourney, it seems.
If this system is better than MidJourney for such, then I'll be busting out these AI children's books much easier. It's a numbers game. Make 5 or more children's books a day with AI, and legally sell them on Amazon, the largest store in the world. Income on autopilot after that. Thousands and thousands of dollars per month, even if each only sells 20 copies a month.
Then I break the Matrix. Hehe. I break systematic poverty.
^_^ I'm part Italian, southern and northern, plus Greek, so, very cool with Leonardo. Hehe.
I've been playing around with this but haven't really gotten results I want, so thanks for making this video!
A really nice quote I heard when looking at anything created by AI was "Remember, what you are getting from an AI at the moment is the worst it will ever be".
So true...and quite scary :P
Nice one, thank you! It is amazing to see how the competition between the ai tools has started now. Only good for us!
Agreed! This is an arms race and consumers win every time.
Leonardo is just what I was looking for, thanks for making this video!
Glad it was helpful! Let me know what else you'd like to see.
@@matthew_berman You have a new sub in me!
@@ucpmod Much appreciated!
Nice one Matthew. Added myself to the waitlist. Looks promising.
Thank you!
so did you get in yet?
@@mikelytou nope not yet. Lol
@@kennethochiabuto8516 oh no. That's a long wait then. Thanks for replying, at least now I know.
@@kennethochiabuto8516 FYI: I got my invite today. I don't know if you did that, but I joined their discord as they suggested and just said hello. Maybe that helped accelerating my access.
oh damn thank you for sharing this! :D going straight for it
Thanks!
Very cool. Can't wait for a self-hosted/local version of this!
It already exists! It's called stable diffusion and you can install it locally.
Here it is: th-cam.com/video/SnYP1EPBIWM/w-d-xo.html
Nice work Matthew, thanks
Thank you!
I must be the only person on earth who hates discord, I despise discord I cannot figure out why any service would want to attach themselves to that interface. Having said that I would probably compare this service to playground AI, it looks like a clone of playground AI.
I don’t hate discord but I’d rather have a web interface for sure!
I'm with you. The interface is a mess. It reminds me to sites like Quora for difficulty to find anything in any semblance of organization.
@@rayfellers you are my people
Discord is awesome. For so many reasons.
Your not the only one
Excellent work 🎉. Keep it up
Thank you, this AI it's fabulous. Your video is great, keep it up ! (Sorry if I make errors, i'm french)
Thank you!
Wouldn't go as far as to say that this is "better".
Midjourney is actually very powerful and extremely precise if you're willing to spend the time training it using seeds and previously generated seeded images, and weighted prompts, to hone the results. You might get lucky, but you probably will only get close to what you're shooting for with a simple text prompt.
I expect Leonardo will improve, but right now the quality of the renders is V3 Midjourney, at best.
They are both based on SD...so they seem pretty comparable to me.
you said Leonardo isn't better and then you proceeded to explain why it's better
@@maurycy_gnc
No I didn't. Right now, the BETA AI Canvas features in Leonardo are really the only thing I use it for. And usually the images I'm modifying are images I generated on Midjourney.
Leonardo AI is way better than Midjourney when it comes to Celebrities, assuming they are well known today.
Midjourney outputs pictures that look kinda like the celebrity you entered,
where as Leonardo Ai is right on, most of the time.
If it’s free, it’s already better in my book.
Certainly some advantages currently, for me personally those include unlimited use (which won't be permanent without paying money) and the 16:9 aspect ratio! Thanks for this video Matt.
thank you! glad you enjoyed it
Midjourney alreadybhas the 16:9 aspect ratio
@@hugomouteira7015 for free? I haven't checked in a few days, last time I did it was only 3:2 for landscape for free trial testing
It should be on the free version too
@@Mattorite i know it was for v3, not sure about v4 though
nice and relaxing explanation
Glad you liked it
i was looking for midjourny alternative , now i got it. therefore i liked your vid and subscribed
thank you!!
Really amazing and practical tutorial, thanks for sharing. Actually I don't like discor niether so this tutorial was really helpful for me.
Glad it was helpful!
Thanks Matthew, signed up for access to the website, just got email to create account etc, looking forward to using it :)
Awesome! They seem to approve accounts very quickly. Thanks for the comment!
@@matthew_berman yes, I only waited around 1 hour or so, loving it so far, best wishes the the AI content, hope it all goes well for you. Again, thanks for showcasing Leonardo AI, appreciated.
Thanks for this great and informative video. This is very helpful
Thank you!
@@matthew_berman :) I just discovered that you have 1000 tokens per day, so 1 per picture. That is already very very good :)
Oh yea? Looks like they _just_ added that limitation. I only got 200 tokens per day haha. It's still plenty.
@@matthew_berman i think it`s new :) yesterday i have no limitation :D
Thank you so much this is great info
You're welcome!
Here's a FIRST LOOK at Adobe's new generative AI product firefly: th-cam.com/video/vtTR8FFvTcM/w-d-xo.html
Also, if you like this video and want more AI tools and tutorials, make sure to subscribe to my substack: matthewberman.substack.com/
"Free" but you are required to make an account?
@@GrimReaperNegi Yes
subscribed , thanks for sharing
Thank you!
Thank you Matthew I appreciate that , would you please elaborate more about the custom database models??
Check out my other video about creating custom models. (Link in description)
thanks for sharing!
You’re welcome!
thanks matt! 👍👍👍
You’re welcome!
Awesome! Thxxx 🍻
You’re welcome. New video on Leonardo coming soon!
Very helpful 👍
Thank you!
Thank you for sharing this interesting video.
Glad you enjoyed it
You used one of my creations at 8:07 😁 believe it or not that was supposed to be a coral reef in a glass, but I used the Nightmare Generator model 😂
One of my fav images. Scary indeed. Thank you!
Thank you 😊 I do enjoy my scary stuff. Awesome video btw 👌 just so you know the DreamShaper 3.2 model does way more than just anime. With enough prompt work you can make some pretty awesome things with it.
Nightmare fuel
Thank you!
You're welcome!
Wow, as a not super techie person this looks a bit intimidating . . . and still fascinating. I guess success comes by getting in and practicing. Thank you for this video. New sub 🙂
Thank you! Best way to learn is to just start :)
Thanks for this, Matthew! I went over and signed up. Is there an estimated amount of time before they will get back to me, or do I simply play the waiting game?
Looks really cool and I can't wait to get in ^_^
I'm not sure, sorry. Best you can do it wait :(
Finally a web interface. Not everyone wants to be part of "the community" or be a zealot of discord
Agreed!
Thanks Matthew for your video I just subscribed to your channel.
Can we use the same prompt used in other tools to be used in this tool as well ?
You can but it'll lead into different results
Definitely, but you will get different results of course. Would be cool to see the differences with the same prompt on different tools.
I'm a total newbie and have a general question. Can you use ai created images you have prompted to illustrate a book
I am using it and it is amazing! I hope they came with resonable pricing plans for casual users like me
I hope so too!
Given how quickly this tech is improving, can you imagine where it's going to be in another 20 years!
I just read the FAQ, and this beats out MidJourney because content belongs to the creator who created it!
I know, it’s pretty incredible. I just got access to Adobe Firefly so going to try that out today.
20 years?! I'm wondering where it'll be in 20 months.
:p MidJourney they say is $10 minimum a month after 25 images. I've also heard it's free as long as you use the public threads. While MidJourney is cool to see others generate art as you do, it's tedious to scroll for and look for yours. It's just easier to have it appear like normal AI does.
I only use MidJourney for the URL link option for making similar art. Lol. Some other AI systems use URL and/or image to reference to, but as I've mentioned, they're not as good as MidJourney, it seems.
If this system is better than MidJourney for such, then I'll be busting out these AI children's books much easier. It's a numbers game. Make 5 or more children's books a day with AI, and legally sell them on Amazon, the largest store in the world. Income on autopilot after that. Thousands and thousands of dollars per month, even if each only sells 20 copies a month.
Then I break the Matrix. Hehe. I break systematic poverty.
@@brokenfingers9607 fr. Or 20 minutes! Hahaha xD 💀💀
:3 I'm part Italian, southern and northern, plus Greek, so, very cool with Leonardo. Hehe.
Question: if I train my own model in leonardo, the images generated are based only on the images I picked or the code picks from other images outside my selection?
It'll be generated based on your custom model.
If you can train the AI to imitate a certain style will there be a problem with copyrighted issue like you use Pokémon's copyrighted images in this tutorial too
Definitely something to consider. It’ll probably be similar to hire copyright law is now, but still some grey area today.
I'm new and wanna start tinkering with AI art and looking for cool ideas.
The big drawback of most prompt to image steps is lack of control. What if you want a slightly different outfit, pose or lighting? I dabbled with DAZ3D years ago, tried blender but found it very focused on experts sculpting stuff and people from scratch.
Is it possible to combine various techniques? You need a basic rigged model, features that can be modified with sliders, add-ons like hair and clothing and textures.
If you have a base model, can you generate these elements with AI to later use as a toolset?
I just watched an video with merging models (and lora) but that seemed for advanced users and still 2D focus. Another video generating 3D (pointcloud?) images seemed promising also, but not that tweakable.
Wow great comment. I think tinkering with the prompt and/or using custom models is the way to get more control. Did you see my video about creating custom models on Leonardo ai?
There's two methods which can be used: Inpainting and ControlNet.
Both are a bit more advanced, but can be done using the self-hosted Automatic1111 interface for Stable Diffusion.
With inpainting, you just use a "marker" to blot out a part of the image, and then you use the settings, denoising & prompt to either add something completely new (throw an airplane into that empty sky!), or to transform the underlying content ("keep the face, but make it 50% more android"). With ControlNet, you can do things like extract a character's pose from one image (or sketch, or 3D model), and apply it to your generated picture.
Thanks for sharing this! I just signed up today....do you know how long the wait period is?
You’re welcome! It seems to be a few days from other comments I’ve been reading.
Thank you
You’re welcome!
i whitelisted 2 days ago but I am not sure how to start?Cani use the prompt in the discord channel or do we need to wait for en email?Any help appreciated
I think you need to wait for the interface. Any luck yet?
With something like this, are there licensing issues on using images you generate for (for example) social media posts etc?
There might be, you have to read through the TOS on any of these tools.
I'm on a waitlist now. Joined their discord but I believe they are moving into a paid model right now. We'll see. Thanks for this info.
Thanks!
did you get in yet? If so, how long did it take?
FYI, 3 days after signing up for the waitlist, I got an invite.
@@mikelytou thanks for sharing!
Hi Again
What do you mean by training the models , more variations?
You can upload your own images and train a model on those images. Hope that helps!
Love your channel pls create one on sinbot
Never heard of sinbot I’ll check it out! Thanks
The artificial intelligence picture is developing day by day.
Thank you for posting the video with good content today.
I'd appreciate it if you could continue to post videos of useful information.
Have a nice day today.😍😍😍
Thank you!!
It's not "completely free and unlimited," you get a certain amount of credits per day -- but they're awfully generous with credits, so it's still a great deal for most uses... I also don't feel like the quality is nearly as good as Midjourney's, but there are a lot of variables that I haven't really messed around with yet, and it's new so there will probably be lots of little improvements coming. That being said, it certainly is a great alternative to paying or subscribing for AI images.
It was free and unlimited when I posted the video, that quickly changed though :\
@@matthew_berman Yeah, I was there for the beta and noticed the change one day, and it looks like the counter still has some bugs -- sometimes it doesn't show the correct number of credits being subtracted, but when I hover my mouse over the number, a more accurate popup appears. I imagine that will be fixed soon... Technology is at the point now where a LOT of things are going to start changing quickly!
@@HyperNova137 The speed at which this entire AI wave is moving is insane. Glad I'm here for it :)
if I try to create with the SD model, the picture is much worse than with the Leonardo diffusion model, why?
I think it depends on the type/style of image you're trying to create. Different models are better for different kinds of images.
Can this be used with Private sessions? As in does every one of my generated images appear publicy? or do I have the option to share them with others?
You can create them privately AND share them publicly if you want.
@@matthew_berman thanks for the reply!
If I may ask,
This is a "2D only" image generator - am I right?
The gray on black UI is the only option available - am I right?
Thanks.
It can do everything. 3D or 2D.
@@matthew_berman Everything would be quite a lot.
Hello,
Seems very interesting and explained nicely. So just to confirm, Leonardo is free to use and we can generate unlimited images?
Also, one more question, let’s say i want to make a children’s story book and want to generate consistent images for my book, how can i do that with this tool? What prompts/methods exactly do we use?
It looks like they literally _just_ added a limitation. You get 200 credits per day, free. Each image is a few credits (I've seen 4 and 5 credits depending on the size).
Here's an example prompt from an image from the video that has multiple angles of the same character: "Detailed, realistic, Half elf woman with flowing brown hair, she is wearing a long embroidered green dress, model sheet turnaround, full color, front view, back view, side view"
It looks like the last keywords "front view, back view, side view" is the part that triggers it.
@@matthew_berman In MJ, we use link of generated image to get rest of images in same style. In Leonardo, I tried to use first image as seed, but results weren't that promising. Did you try or any idea to generate same image style with different angles?
I'm also looking for images for storybook. If you get better result with Leonardo, please share. I'll update too.
just signed up thanks
You’re welcome!
I love that paper art style. Is there a way to download that model, I'd use it on my SD1111 web ui? P.S. I have few ones from civitai, but they are not producing high quality images. P.P.S. Good video bro.
Thank you!
I’ve never heard of SD1111, can you explain? I don’t believe there’s a way to download the model. That might be bad for their business ;)
Any detailed pros and cons available for this vs Mid Journey? As it seems both are using their own proprietary AI tech I'm really interested to know of any advantages one has over the other, or disadvantages. I dont care so much about the user interface, I mean more about what kind of image can be produced from a given input prompt.
Both are using stable diffusion, so the results should be similar in terms of quality.
@@matthew_berman I can get very similar results witj MJ and SD but casual user cannot. MJ has training wheels on and they have made very hard to make a bad image (they limited negative prompts just recently). Use: 32307 as a prompt in both to see how differently they react to single token.
This token usually produces at least one similar image in both: worldenvironmentday
@@digidope Interesting. Do you think a video comparing the results of both would be valuable?
@@matthew_berman I think a video where someone manages to get nearly identical image from MJ and vanilla SD 1.5 with just prompting would be interesting. There was a way to "break" MJ from the MJ look with negative prompt but they have limited that in recent updates.
@@matthew_berman Midjourney is not Stable Diffusion. Midjourney, Stable Diffusion, DALL-E, are all different AI, created by different people (Midjourney by Midjourney Inc., and SD and DALL-E by OpenAI, the same people who created ChatGPT). Also, Leonardo is not an AI, it's just a wrapper for Stable Diffusion, and possibly DALL-E, I haven't checked, like, for example, NightCafe Studio is, and probably many more who are merely using someone else's AI (Stable Diffusion and DALL-E).
Stable Diffusion is not a substitute for Midjourney any more than a PC is a substitute for a Mac. They're two different products.
I already use DaVinci for editing, so Leonardo will fit in nicely.
Can you elaborate? DaVinci is video editing, how will an AI art tool help? Would love to hear more.
@@matthew_berman :D both are tools that will help with expressing creativity, but together they are Leonardo da Vinci. That's the joke.
Nice Vid ,when i click on it it takes me to discord how do i get the site interface you have ? or have they changed it to discord only?
You have to sign up, get in the discord, and then wait for an invite. I’ve heard it takes a few days.
so you register for early access, then verify your email account, then it refuses to let you create an account to use the service? Or is there some IQ prerequisite I'm failing to meet in order to use this?
Just gotta wait :(
Is there any countries where Leonard ai doesn’t work ?i tried many times but the application doesn’t work .i made every step requested .i messed somewhere and it is frustrat
Not that I’m aware of. What error are you seeing?
@@matthew_berman thank you for your answer.actually my iPad is an old guy.i got Leonardo ai on my iPhone successfully.but the screen is tiny not ideal to create .anyway thank you for your answer .most success and prosperity
@@toffahahloua6781 thank you!
welp, it's giving me plenty of ideas thanks
Thank you!
Thx for this great tip.
What is the max image resolution we can get compared to Midjourney ?
And I guess it might be free because it’s at the beta-stage ?
I can’t remember the max res but I think 1024x1024. Thank you!
@@matthew_berman Wow not much. No way to print that… Would you have by any chance a recommendation for a good app to enhance the resolution 😳 ?
@@betterburnout Absolutely. Check out this video: th-cam.com/video/pG3JbtF-mNo/w-d-xo.html
@@matthew_berman Thank you ! I'll check it ✨
@@betterburnout You're welcome!
Thank you so much, but I can't see same main menu Like you? how can I go to home and start here?
Hmm...what are you seeing?
Do you know how long it usually takes for them to allow access to the website?
Took me less than an hour. Thanks.
Ohhh Ive got a few questions. 1) are the images Generated Unique or are they Taken from Bits and Pieces of other art? 2) WTH do they make it so hard to sign up? Ive literally been waiting 40 mins just to be "waitlisted"?
The images are unique in that they don't exist elsewhere, but all AI is trained on previous data. So any generative AI that you see is a derivative of something it has seen before. Sorry about the signup process, it does suck. It use to be easy and no waitlist.
Sir, I have a doubt, I want to pay a month just to test Leonardo there, is his payment renewed automatically? or do you really pay only one month
I don’t know if they have a paid service yet.
I'd love to try it out myself, but they aren't giving out any more invites. So I have no idea if it's truly better than midjourney. At least with midjourney & blue willow, I could create images from day 1. I didn't have to give out personal information & wait, hoping for some magical invitation.
Do they explicitly say they aren't giving out more invites? It has been "wait" not "no more".
@@matthew_berman" Due to the vast surge in demand of registrants on our waiting list for an Early Access invitation and as we are still currently working hard to optimize our infrastructure and service, we are currently unable to provide an estimate on when invitations will be sent out."
@@Zeoran Yea...ugh sorry :(
there is a limit now of 25 images I think now, and still waiting to join as they have been overwhelmed by the number of users joining
I believe it’s a credit system, 200 per day. 5 credits max per image is what I’ve seen. So more than 25.
While I don't agree that the quality is as good as MidJourney's, the interface seems neat. Also, whenever I come across a new ai software like this, apart from the end result and workflow, the first thing I look for is an option to work in private. If I have to do client-related work, there's no way I'm letting everyone see or download the WIP.
I've found the quality to be on par at least with MidJourney, but everyone's experience will be different.
As for working in private, both MidJourney and Leonardo both allow for that. MidJourney you can private message the bot and Leonardo you just create without sharing.
Though technically even when creating in private messages with MJ, the image will show up on your profile on the website and potentially on the community feed if it gets voted up.
MJ does have a fully private option but it requires paying extra.
@@ShawnFumo exactly, and everyone I know who uses MJ for client work gladly pays that extra.
@@matthew_berman I disagree, I reckon its really nowhere near on par with Midjourney. Where this scores is that it is a usable interface for low end use of Stable Diffusion thats integrating some of the better open source features, but its weakness is that to get really decent results with Stable Diffusion you need some petty serious GPUs. If you want the best results you need to run it with a GPU worth thousands. The folk I know doing that generally rent them. Midjourney is not *just* Stable Diffusion btw, it has its own secret sauce. which is why it has its own version releases.
I tried to open the website for curiosity, it says that is an early alpha, am I wrong?
And is there something for Android? I would like to us it with my phone.
It’s web based, should work on any device
Thanks for the video, but when I searched for Leonardo, and joined, it's part of Discord, just like the way Midjourney is.
It might have a discord channel but it definitely has a web interface as you can see in the video. Let me know if you aren’t able to find it and I will try to help.
@@matthew_berman Thank you, I'm in
Oh No! should of kept this a secret 😁 been on the whitelist for a while, Leonardo is great
Haha happy to share it with others!
leonardo is also new- so its free for now but i'd imagine a pay model is coming if it doesn't already exist. But since i can't see it who knows... :(
Yes definitely going to be paid
MidJourney still top of the line when it comes to AI generated images, depending upon how realistic you want the image to look.
So is this just another stable diffusion interface?
You make it sound so simple ;)
Yup - but it looks like a petty well thought out one, including some of the latest and most useful open source features like Dreambooth and Controlnet. If you are looking for a good front end to run a low GPU powered version of Stable Diffusion with very little skill or effort or tech know how, this is better than most ive seen. If v5 of Midjourney included these features it would blow my mind, but then Midjourney isnt *just* stable Diffusion, so I think its harder to bolt these open source models in. I studied user interface design at uni as part of my computing degree and this one is really pretty good. A few clunky bits that are too hidden away that need to be more obvious, but theres a lot of usability there once you get the hang of it.
I'd have been more impressed had you explained what the hell all those acronyms were.
Which acronyms?
@@matthew_berman Hi Matthew - thanks for the reply, not everybody can be bothered to answer. For a start, "RPG", "FPS" and "UI", (User Interface?), as shown at 3.48.
Additionally, whatever is charizard? What I would *really* appreciate is a video showing how to use Leonardo for people like me, who know very little about how to generate AI graphics, please. I'd be willing to bet there'd be others who'd appreciate that too.
Many thanks and subscribed.
The main thing that attracted me is … Free … guess that ship has sailed 😢
I think it’s still free, just not unlimited
thanks! :) can you use your creations commercially?
I replied to another comment with the same question. The answer is many places allow you to sell AI generated art (like adobe) and most tools give you the rights to your creations. Thanks.
Yes you can.
@@leonardoai thank you! :)
After messing around with Leonardo for a good hour or two, I can say confidently that the quality VS Midjourney's V4 and V5 doesn't even meet a fraction of the expected results. MJ does exceptional backgrounds and has it's own driven styles. Leonardo however has a different sense of aesthetic entirely but often felt plasticy and uninspired. I did however find it generated good simple icon-type assets, but nothing I would actually use it for.
It's a case of weighing the pro's and con's for your needs, though Leonardo does has a more appealing basic subscription plan package, until it can catch up with the quality I've come to expect from MJ, I'm personally going to give it a hard pass. Maybe in the near future I'll switch.
Another personal thing, I actually really like the MJ discord bot. I can use it on the fly and don't have to deal with websites. I can just drop my promp and get on with my day while out and about.
Leonardo's browser UI felt clunky to me and just to get a simple render there was an awful lot clicking about to find the right place. I would hate to imagine how bad it would be on a mobile - I didn't test that.
Fair enough. I still prefer the web UI though.
Leonardo AI has huge problem with multiple head, multiple bodies and distorted arms and legs.
about 4/5 of full body picture generated with Leonardo seems to have some form of disfigured boy.
If they can improve this, this would be the best tool.
I’m sure they will iterate and improve. Thanks for the comment!
Please make a video on how to show two characters in one scene fighting each other....I want to make an action scene in which two or more characters are fighting each other but midjuorney and leonardo are not able to do this when I try...each time it make scene without action or fighting... please make a video on action scene of multiple characters in one scene
Probably won't have time to make a video on this, but thanks for the request!
the quality is already a hundred times superior than stable diffusion, how i never heard of leonardo before???
It uses stable diffusion models, so nothing new.
@@maxdefire i've been using stable difussion locally on my pc using trained models from civitai, and i'm getting some impressive results comparable with leonardo, the thing is, i haven't found a trained model for assets
I don't get the same page as you do? I even have to download a discord app? I'm so confuse. Totally new in this
When I recorded the video, Leonardo was open access no limitations. Now you have to wait a few days.
Hi Mathew
Where can I download this?
Leonardo.ai
does leonardo not work on a token system?
It does last time I checked.
could u plz tell me how to download this software plz provide the link
There’s nothing to download. It’s all web-based.
except its not free there is a waitlist "we are currently unable to provide an estimate on when invitations will be sent out."so who knows when i can actually use it :(
It's still free, there's just a waitlist.
Hi, The site does not let me create an account, It says I should be invited.
Can you help with this?
I wish I could help. Is it invite only now? Are you not able to get on a wait list?
@@matthew_berman I could but I wanted to try it now
@@matthew_berman Is it possible for you to invite me?
@@alirezazahedirad4847 unfortunately there’s no way for me to invite people
How do you create or train AI models ?
I’m putting together a video about that right now! Should be ready in the next day or two. Come back soon :)
Do you get commercial rights to your images?
Yes 🙂
Are you the creator of Leonardo? I absolutely love your product. Where can I find the terms of use?
@@leonardoai is it using stable diffusion backend?
So this is online only? I am looking for offline but I will give this a whirl to know what I like. Thanks!
Offline doesn't work unless you download the entire thing and have a very powerful computer.
@@ImARealHumanPerson I
have a flavor of stable diffusion that I can eek out 384x384 with 20 runs and batch of 1 on an AMD RX6600 8Gb.
So this isn't currently available to everyone. When is that happening?
I'm not sure, it was available to everyone when I posted this video.
@@matthew_berman There was a waiting list, I just got added to their whitelist 3/16/23. The free version limits how many images/upscales you can do in a 24 hour period, if you want more tokens for the month you have to pay.
Been using Leonardo since yesterday, very nice indeed but my question is what happens when we used our 250 tokens? Do we buy? Do we get tokens at the end of the month? If someone can answer.
It resets each day and you get more tokens.
@@matthew_berman Thank you for the quick answer!
I'm part Italian, southern and northern, plus Greek, so, very cool with Leonardo. Hehe.
Vegan is hard to stick to, but I appreciate it!
wondering if you can sell some of the art you generate here???
I believe so but check the ToS just in case
This tool is not yet available publically. You have to join the waitlist for this.
It was. But they closed it down because it got so popular.
I have been gathering a "crew" of sorts that I can rely on for different specialties. So why not add this one to it.
Cool :)