Fantastic AI Image Platform That Will Level You Up! (FREE Credits!)

[@TheoreticallyMedia]

Claim your FREE credits in the description link! As always, I'll keep an eye out for more free stuff for you all!

[@neoced9293]

how many picture can you generate with one credit ?

[@neoced9293]

how many pictures can you generate with 1 credit ?

[@livinghistoryaienhanced]

Free credits via your affiliate link is not working. I didn't get the extra 50.

[@nicheva417]

As a user, two things I would like made a bit easier are 1. Easier consistent character across multiple generations. Ability to overlay text (like comic book speech bubbles and story book text). In fact, just an easy to use story book template (page 1, page 2, turn page etc) that can be exported would be great. Or have a one-stop landing page for user to create storybook/comic etc. Draft and publish phase. They have so many advanced settings I feel like this should be one of the easier functions. I think t demand would be surprisingly large.

[@TheoreticallyMedia]

This is a FANTASTIC suggestion! I'll put it top of list!

[@nicheva417]

@@TheoreticallyMedia The only thing I would add, is you want to publish page by page. Most people wont be able to complete an entire "book" - but if you can just do a page from time to time, I think there is increased chance of sustained incentive and habit building.

[@coco-openart]

Thanks for the suggestion! This is Coco from OpenArt. +1 with Tim that this is FANTASTIC! By any chance that you may have a story book example online that you can share that showcases what you want? Thanks!

[@coco-openart]

Love this suggestion! This is Coco from OpenArt. Is there an example online you could point out about the final book you have in mind? Thanks!

[@harrygoldhagen2732]

You can drop your images into comic life 3 for speech bubbles, frames, layouts, and printing page by page. Inexpensive app!

[@JeanPhilippeMarsh]

Hi there. Very useful intro video for newbies. I actually started a web design business last month and my first client was a thriller author that wanted me to create a web site with AI generated images of his books. Boy did I have my fix in delving head first in AI image generation! He is so wowed that he keeps feeding me projects long after the site is delivered! Now you've got me VERY interested in becoming more efficient. 🤩 Feature request (I'm not the first one, obviously): character consistency ported across continuous images. Thanks much!

[@TheoreticallyMedia]

Character Consistency is kind of the holy grail. It’s getting there! Just talked to the OpenArt crew and that is going to the top of their list!

[@coco-openart]

Congrats on the new AI projects. We're actively working on consistency character and we hope to have something amazing in the next 1 or 2 months :)

[@justinlloyd3]

I'll tell you what I want in an image generator. Difference prompting. Or maybe call it directional promting. I want to be able to say "this hat is too green" or "Move her more to screen left", or "remove the orange lights". Currently you just say what you want and it overwrites what you just did, this is the case for editing like in mid journey. I want to be able to talk about what was just made and indicate a direction to go next.

[@coco-openart]

Thanks @justinlloyd3! This is Coco from OpenArt. This makes a lot of sense. We'll explore in this direction too!

[@TheoreticallyMedia]

That’s a great note. I was just talking about that on the video generation side. Like, chatting with a video, as if you were a director working with an actor and a crew. …and also laughing at the idea of an AI actor arguing with me about their motivation! Haha. Anyhow, yeah, what you’re talking about is exactly what ChatGPT and Dalle3 implied in their previews. It did not happen, obviously- but it’s such a great idea.

[@ronvanderlee7298]

Absolutely this. To me, this would take AI image generation to the next level. At this point with mostly regeneration of the whole image, it's very much a role-the-dice exercise, and burning credits to boot. With direction and specific objects targeting it would be so much more useful. Just how useful everyone on a set is; dop, gaffer, director, etc.

[@shehabdentalclinic857]

Yes, great suggestion. Playground applied that in their image editing directional prompting. This lower the barrier for the end user in editing with natural language, and makes a much easier user experience interface.

[@jeremyrobinson9082]

Awesome. Finished a novel. screenplay and graphic novel script this week, so I'm using today and tomorrow for R&D. Perfect timing!

[@TheoreticallyMedia]

You have been BUSY! It’s funny, I actually want to take a deeper dive here and try out a mini graphic novel tutorial with OpenArt. Might be a good spot for you to work on that script!

[@jeremyrobinson9082]

@@TheoreticallyMedia Lol. So busy. My personal theory with AI art is that I try to only replace myself. In addition to all the writing, I'm also a book cover designer (previously for other people and publishers, now just for myself). I make all my graphics, trailers, promotions myself and I'm just one person so AI art, voice, music, video, makes my life easier. Since I'm NOT also drawing comics (though I used to do that too!!) I don't use AI for my comics. I'm currently contracting 7 real artists/colorists to bring the comics to life. That said, I am interested in short film via AI, and perhaps a very strange comic that I wouldn't release in stores. But again, busy already! Until AI can collaborate with me through an AGI or something, I'm still limited by human time.

[@VHSTrailerPark]

@@TheoreticallyMedia Yes, please! A tutorial on creating a graphic novel with OpenArt would be great! I was searching for something like that all week. I'm a screenwriter and want to transform some of my scripts into a graphic novel. My problem is I do suck at drawing, haha! So if I could use OpenArt to do that for me then I finally might be able to bring these stories to life.

[@jeremyrobinson9082]

As for Open Art, I'm thrilled to see the layers option, but I'd like to see a layers option where you can add text layers and then save as separate TRANSPARENT PNGs. Would make using this stuff professionally much easier. Right now, I lot of Photoshop is required for most images, especially book covers.

[@TheoreticallyMedia]

Great comment! I’ll pass that one along to them!

[@coco-openart]

Thanks Jeremy for this feedback! This is Coco from OpenArt! Totally agree with you. Believe it or not, the moment I'm typing, our engineer is working on the text feature, so it's coming in the next few days. If you don't me asking, why "save as separate transparent pngs"? How will you use them? Would really appreciate more details. Thanks!

[@jeremyrobinson9082]

@@coco-openart Wow, KZbin really doesn't want me sharing an example from my website. It's blocked my response twice now. So I'll be extra sneaky this time. Separate PNGs are important--especially with text layers--because you might need to use the art, or text, outside of the original format. For example, I design book covers, once upon a time for publishers, and now just for my novels. When I promote that novel, I need to be able to use the book's logo separate from the cover, on t-shirts, mugs, posters, as well as in marketing. Facebook cover images, trailers, banners, email headers, etc... A good example of this is on my website, which you can find by googling Jeremy Robinson. Beware of Monsters is what you're looking for. :) The banner at the top of the first post is made from AI art, the book cover, and then a separate use of the logo. So, if the text layer is merged with the art its basically unusable for much beyond a web post. And right now, no one does this, and no one does PNGs with transparency. Right now, MidJourney is the go-to for art that's good enough for a cover, but this would put you neck and neck, I think. I also currently use ComfyUI and have all the models on your site, BUT its so much more user friendly than Comfy. I like where you're heading.

[@jeremyrobinson9082]

@@coco-openart It appears that my last three attempts to reply to this have all been removed by YT for some reason. Okay if I message you on Discord?

[@thehippie3610]

saving individual PNGs will let people create and save scenes and pieces of scenes.​@@coco-openart Save the trees in the background, but redo the cabin in the foreground. Allowing selective generation of the individual layers, so you could "save" layers 1-3 but regenerate the layer 4 while keeping everything else you like.

[@harrygoldhagen2732]

This is all great, Tim, but the key feature for me is consistent characters that you can pose. The holy grail!

[@TheoreticallyMedia]

Getting there! I'll talk to them about pushing the envelope here a little more! I TOTALLY agree!

[@TheBARONEARLSShow]

Easy Consistent Character is the #1 thing I want & hear other creators asking for. That's what will help us tell stories with our characters. I would also add Consistent Characters across illustration styles would be HUGE as well

[@TheoreticallyMedia]

10000% agreed. There are some consistent character hacks, and some model training- but it often is still a lot of work to get a reliable output. I know OpenArt is putting this at the top of their list though. So, stay tuned!

[@TheBARONEARLSShow]

Thx for all you do, man. I know AI makes some things easier, but it's still a hell of a LOT of work to do a video podcast at your quality level. I’m a disabled pro creative, so I explore any tool I can use as an “artificial limb” to help me continue creating art, music, design & stories. Your show is empowering 🙏🏼

[@johndev2212]

The precise upscale function is a very welcomed addition.

[@TheoreticallyMedia]

Right? That’s the first time I’ve seen that outside of the traditional upscale world (like Topaz). But on that side, you don’t get the creative upscale. Nice to see options for both!

[@mikemckenna4778]

Greeting! As a filmmaker and photographer, I am just making the leap into the AI medium, so your channel has been a great watch. I am currently sampling Midjourney, but am open to any platform that best fits my needs. And that would be to create visual shot lists and storyboards to use as prompts for AI filmmaking. So, character consistency is crucial.

[@TheoreticallyMedia]

Excellent to hear! And don’t get me wrong: I still LOVE Midjourney! To me, it’s always, what’s the right tool for the job? But, I’ll say: kitbashing and using multiple tools can be a powerful level up. Say, taking a MJ output and using OpenArt to edit it. As much as I love MJ, currently, the editing tools leave a lot to be desired.

[@seanmaynard7417]

I have been using openart for image generation and fixing images. It is a good added utility. I like the live portrait option but it could be a little better with the results at times. One of the best things I like is fixing badly created faces from other image generators.

[@TheoreticallyMedia]

I think you nailed it, it’s a great utility platform. But, once that Flux Dev model is up and running, it’ll be a powerhouse, And I know most don’t dive into the Comfy side of things, but the fact that they’ve got it available for use? That’s just awesome. Can’t think of anyone else doing that!

[@coco-openart]

Thanks for your kind comment Sean! We will keep improving the results of the facial expression and making it more real time too.

[@richardglady3009]

Great introduction. Thanks for the link giving more credits to experiment with. Thanks for the video.

[@Deb____]

Thanks for the free credits and since I'm brand new here, thanks for all the awesome help!

[@MetaPixArt]

I enjoy your KZbin Thumbnails! They’re simple and authentic which stands out a lot.

[@heartshinemusic]

I saw briefly that you can also create your own "Character Model", would be interesting if you could maybe show those results. When combining that with the "pose" reference feature, you could get a lot of control. You could even use those results as the input for a new image to video experiment. Thanks again for all the updates on new A.I. image/video stuff!

[@TheoreticallyMedia]

Yeah, sorry! There was a LOT to cover in this video, so I had to make some choices in terms of what to cover with a little more depth, and what to briefly mention. Great note thought. Might make for a good follow up video!

[@heartshinemusic]

@@TheoreticallyMedia Don't be sorry, Tim. You're doing such a good job already!

[@BakerImageGroup]

Would be nice to be able to set keyframes for the poses then export as a sequence.

[@TheoreticallyMedia]

They have this interesting bulk generator, where you can prompt and have like 30 outputs generated at one time. I didn’t cover it here, but your thought is exactly what I was thinking when I saw it. If there were a way to use it as a keyframer? That’d be amazing.

[@csebastian71]

Since I dumped Adobe months ago, been waiting for an alternative program to edit photos and create timelapses from photos. This looks like a good place to start and dip my toes in. Thank you

[@dianegrant3215]

Well, I have been watching all kinds of different KZbin videos, especially Matt Wolf, this is the first time I have seen an application that I actually would like to play around in and as a beginner I’m super grateful for the tutorials, and enthusiastic about giving this a try! More comments to come as I gain my experience in here. Thanks so much.

[@TheoreticallyMedia]

that's awesome to hear! Love Matt's channel as well! If you're just getting started, for sure try out the Flux model! Once they get the dev version in October, it'll be a real powerhouse. In the meantime, lots of other cool models to explore! And then there's the whole Comfy UI side, but: one step at a time!

[@High-Tech-Geek]

Thanks Tim! Looks like a very comprehensive image suite. But... I HATE temporary credits that expire after 7 days. I add this site to my list of 50 AI sites and when I find I might want to test one of their features, my credits are long expired. WHY do some sites do this?! It feels so cheap and greedy. Just give us the free credits and let us decide how and when to use them whenever we have the need. Or, better yet, give us the 100 free credits (or whatever) per month like so many other AI tools. Anyone that's allowed to dabble and learns to fall in love with one or more of your tools would then become a permanent subscriber. Let us dabblers dabble. We are not going to convert to full time users unless we can thoroughly test and play with your tools over time. We then support good tools that are useful to us. That's how you convert all those free users to subscribers.

[@Apuat]

I don’t know if this would be possible, but given all of the features it has it would be nice if there was a workflow which was a competitor to LTX studio, (considering its pricing structure), where you could create storyboards or episodic story in various styles.

[@TheoreticallyMedia]

Yeah, I have yet to see anything quite like LTX Studio! They're pretty much the only one with that much stuff in one place! It's pretty impressive!

[@liviere1]

Thank you so much Tim for bring this bad ass to the table of AI, thrilled to start playing/working with it! also thanks for the + credits already join! It has many features that looks very interesting but also may reduce steps in retouching with Ps. Love your content! ✌

[@biomajor099]

I would love it to be able to generate high quality pixel art for use in video games. Like make sprite sheets and ability to remember characters and be able to generate multiple images of the same character for different animations like running, jumping, shooting etc.

[@A-Gambino]

Thanks really concise introduction to openart. Although I've been a member since the early days of openart I thought I would never venture down the Comfy UI road, but after watching your video I'm going to go explore that side of the site as it doesn't seem as mind boggling as I thought it would be so thanks for that ✌🏻

[@TheoreticallyMedia]

Y'know, Comfy is one of those things where-- you see it at first and you're like "NO WAY"-- and then you learn a few things and suddenly a lightbulb goes off and you're like: Oh, I get this. And then you see some Comfy Wizardry workflow and you're like: OH, this can get crazy really quick! But: One day at a time, and I bet you can get to a really solid place in less than a few weeks! It opens up a whole new world-- so, I totally encourage you to hop into it! (FYI: I am NOT a Comfy Wizard at all!)

[@A-Gambino]

Thanks for your feedback I'm glad to see it's yet another new world for me to explore in this digital age

[@Therealfury9]

Can't wait to try it. This looks amazing

[@TheoreticallyMedia]

There’s a lot of cool stuff to explore here! Grab those free credits for sure!

[@jonanderson5809]

Multiple descriptive characters. Every AI I've played with blends characters if you describe them separately. Sure you can prompt, 3 people standing on the beach watching the sunset and the AI will generate that, but when you prompt descriptions of each person, the AIs tend to get lost and you end up with 1 person with traits of all, or clones.

[@cjplay2]

One thing I am always looking for is improving existing older videos. When I finally get to spin up comfyui, that will be the workflow set I want to hit. 24-30, SD to UHD...

[@TheoreticallyMedia]

Oh, I hear you. Currently I think the only thing that’s doing that at a super proficient level is Topaz Video. But it is kinda costly. I have been pretty interested in Krea’s video Upscaler, which has a neat trick where it creates a stylize pass over your video. It’s a little hit and miss, and needs better controls, but for the right use case, it can produce some super cool outputs. But, yeah: it also changes the video, which most of the time, we don’t want.

[@lumideo]

Thanks I've claimed your free credits offer. Let's see what I can learn and create 😊

[@jsa-z1722]

Only ONE thing- consistent characters!

[@coco-openart]

Yes! Completely agree. This is Coco from OpenArt. May I ask a follow-up question? In your case, what are you using the consistent character for? Book illustration or? Thanks!

[@slackerpope]

​@coco-openart Book illustrations and graphic novels. No character consistency no subscription. It's just that simple. Best of luck to you!

[@harrygoldhagen2732]

Comic strips, comic books. Consistent characters is the holy grail!

[@gaelpiratelife1941]

That is easy take 10 20 pictures and train a lora

[@adarwinterdror7245]

​@@coco-openart Im using AI images for video generators, and i create videos that tell stories. Without a way to insert character reference and get the same character in different images - all of these amazing tools just dont matter. If there was a way to create an image with multiple characters and then mask certain character and then insert a character reference in the UI to get the character you want instead of the character that's already existing in the image- that would have been great! In theory, if the above description was possible, this could have been done with other characters on the same image and the user would slowly could change existing characters to the ones he desires. If it was possible to mask a certain character and then use the post reference to change the pose of an existing character in combine that with character reference - that would have been a game changer and that would make what every tool that makes this possible have an incredible advantage overall other tools.

[@FANTASTICWONDERFUL]

🔥Looks awesome! I can't wait to test it out!

[@FewFew77]

This looks like a Midjourney killer. I love that you can pose characters or have your sketches rendered.

[@ian1957ruth]

I am very new to AI image generation. So to test the waters I took a months subscription to Midjourney. About two weeks into my subscription I was told I was banned. I contacted Midjourney who said they do not discuss individual cases, the ban is permanent and there are no refunds. So I did not know the reason for my ban, I still do not know the reason for my ban. So yes for me this is a Midjourney killer. BTW I did not try to create NSFW content, well knowingly at least.

[@TeemuKetola]

For my AI short films, I really need the image generator to be able to create consistent characters. Open Art certainly seems legit, but can it do consistent characters?

[@TheoreticallyMedia]

Kinnnnnnda. But I mean, that's true of all of them right now. We have some face swapping stuff, but as I'm sure you've seen, outfits will also change. You CAN train images here for your own model, but currently I think that output works best for illustrative images. When they get Flux Dev (in Oct) I think we can move into the photorealistic looks. I'm going to talk to them about some new things I've seen that will maintain consistent clothes as well. That's the big thing with this comment section: We've got an open door to a developer to tell them what we want. Hammer away!

[@hyperbolicfilms]

As someone who fights with my local comfy install constantly, this seems like a great deal.

[@coco-openart]

Thanks @hyperbolicfilms. This is Coco from OpenArt. We'll also introduce more ComfyUI dev tools in the next few months. Stay tuned :)

[@TheoreticallyMedia]

That’s awesome to hear.

[@TeddyLeppard]

So many good new companies to investigate now. Fantastic.

[@rotimiadeniyiawaye5950]

I'll give it a try. Especially for the tutorials part...then of course the free credits

[@raxonrex9102]

Thank you. Just claimed my credit.

[@coco-openart]

Thanks @raxonrex9102! This is Coco from OpenArt. Hope you enjoy OpenArt!

[@Mega_Evolution_]

Love this series every time, Tim! 😍Wonder what the next episode will be like though. Can't wait to see! It would be great if you could check your inbox by the way. We've sent you one!🤗

[@legacylee]

funny, I remember using Open Art in their early days... guess i'll swing by to play with the newest version.

[@armondtanz]

We need a showdown between this and freepik 👌

[@TheoreticallyMedia]

Yeah- I think once Flux Dev is everywhere, it’ll really be a matter of what other stuff they offer. The fact that OpenArtAI is running cloud Comfy to use and learn? That’s pretty awesome

[@armondtanz]

@@TheoreticallyMedia lol... why oh why they never called it comfy cloud?? 🤣

[@puja1985]

At 9:20 , does the live portrait expression change in real-time as we tweak the settings? I can't see clearly if it does from the video

[@TheoreticallyMedia]

When I was using it, it wasn’t SUPER real time, like we saw in the demo video the other day. But it would update as you made changes. I do suspect that demo video might have been sped up, now that I’ve played with it.

[@coco-openart]

Thanks @puja1985 for the great question! This is Coco from OpenArt. As Tim mentioned, currently it wasn't super real time. I would say that it would 3-10 seconds. However, we're working diligently to make it a real-time experience. We appreciate your feedback. Motivated us more :)

[@Leto2ndAtreides]

The main thing I've needed is good AI inpainting. Playground for example, has not served me well. And Midjourney doesn't let you upload your own images for editing.

[@TheoreticallyMedia]

For sure give this a shot in that case. This might be the most robust editor I’ve seen yet,

[@Leto2ndAtreides]

@@TheoreticallyMedia Yeah, I just tested it.The results are similar to Playground... Although I watched the tutorial and it does have some interesting things going. But, if we look at their inpainting tutorial... A lot of that looks like you could use AI to automate it, and simplify the process of getting really good images with a text prompt of desired changes instead.

[@coco-openart]

Thanks @Leto2ndAtreides for your feedback. This is Coco from OpenArt. Our inpainting actually has 5 modes. By default it's quick, but I highly recommend the Quality mode too. However, we're constantly improving the quality of inpainting so you can stay assured that our inpainting will get better and better over time :) I also like your idea of automating it.

[@Leto2ndAtreides]

@@coco-openart Thanks. I'll hopefully check at some point.

[@sipunteun]

thank you, for the credit, the restart tips not working btw, some thing need to be adjusted before run, it's better to download it and run locally, it's great platform and happy to learn something new😁

[@Deckard76]

Great walkthrough of this platform! Have you used Invoke?

[@TheoreticallyMedia]

I have! But early on. I need to swing back over there, I’m sure it has changed a lot!

[@faduartist101]

thanks for the free credits

[@steviemac1737]

One rainy day I'm going to keyframe all your thumbnails 😂

[@TheoreticallyMedia]

Haha. Please send it to me when you do!

[@steviemac1737]

@@TheoreticallyMedia 🫡

[@Mrim86]

This looks super powerful

[@ShuangLiu-w6d]

cool !

[@onlineispections]

Hi, thanks for the video, I wanted to ask you if the runway platform in the next version, will introduce character to video, As you rightly showed here in vidu.

[@TheoreticallyMedia]

Not sure-- in the latest video (Massive Week) I show some stuff coming out of the Runway Research Lab, but I didn't see anything on custom characters. I'm bugging Minimax about implementing this feature, and OpenArt is for sure working on it (for stills) based off the feedback here! So, we're on our way! Any discussion I have with an AI Generator, I'm putting Consistent Characters at the top of the list!

[@onlineispections]

@@TheoreticallyMedia Thank you so much for your collaboration it is much faster vidu in creating the character from the face photograph,, Rather than image to video. Pixverse It does this by creating a character but it is not as faithful as Rumway which I think remains the best tool with gen 3. Otherwise you need to create a lora model and insert it into stable diffusion or fooocus but it is not completely faithful. Studio LTX He does it but it's not fraudulent....two or more characters etc. Thanks

[@mr.terpenes578]

lets go man, i miss all your post

[@TheoreticallyMedia]

You got first comment!! Chicken Dinner to YOU!!

[@peacetoall1858]

Looks great. How many credits per image? What's the speed like after the 'Turbo' generations are over? How many credits to create image to video with comfyui?

[@andrewcampbell8938]

If I buy a yearly subscription can I use the cloud-based ComfyUI workflows for free? Are these workflows included in the sub and does image gen using the workflow cost credits? thx

[@TheoreticallyMedia]

Great questions! I....don't know? I'll ping Coco from OpenArt to check in on this question and give you an answer! Sorry, that part isn't my department!

[@coco-openart]

Hi Andrew, thanks for your great question. The cloud-based ComfyUI workflows are free, you don't need to pay for it! There're however 2 limitations: 1. Not all the workflows are cloud runnable. We only deployed some good ones to be runnable. 2. The session expires within an hour. We might offer something better cloud runs in Sep and Oct! Thanks for your interest

[@DataJuggler]

One day in the near future, a new AI Image Generator will be like a new Starbucks or a new Mattress Firm.

[@TheoreticallyMedia]

I'm still amazed at how quickly a Starbucks can...uhhh, generate? I swear, I saw an empty lot the other day, and a week later: Starbucks with a drivethru-- already packed. How do they do it?!

[@Thedarkbunnyrabbit]

If they could use Novelai's custom SDXL model it would be great. $14 a month isn't the worst price ever, but I have no idea how many image 500 credits makes. It looks like maybe 50? On top of that, their NSFW tier is $26 for some reason, and the unlimited one (which is really a must for any serious use of the models) is $56. It's clearly priced for annual, but I aint got the money for annual subscriptions.

[@TheoreticallyMedia]

Good note re: pricing. I’ll give them that note, along with the custom SDXL model as well!

[@Thedarkbunnyrabbit]

@@TheoreticallyMedia If they do end up changing the pricing structure a bit - especially if NAI is willing to share model access with them - I'll be quite interested in the site. I will keep an eye on it.

[@coco-openart]

Hi there, this is Coco from OpenArt. For the starter, we actually offer 5000 credits, not 500 credits :) Though we don't offer Novelia's custom SDXL model, we do have many other models, many anime related. Check out them :) Thanks!

[@bm63]

How do credits work I had a quick look but maybe I missed the explanation Obviously credits are used for certain things but what and how…. Credits kind of put me off.

[@TheoreticallyMedia]

Good note. I was obviously playing around with the middle tier plan, but to be honest- when I’m working on these types of videos, I’m not focusing on credit cost- more focusing on the technical aspects. But, that’s obviously a super valid concern for the actual user. I’ll pass them this note re: credits system. It’s funny, as I’m thinking about it- like, Midjourney had their “hours” model. You can do sort of anything you want, but it’s based on “hours”- yet, I don’t think anyone has ever questioned what part of an “hour” a generation costs. Huh…well, I guess that’s my “shower thought” of the day!

[@shadowfall9561]

I usually use Leonardo so how does this stack up against Leonardo? I like that you can use different models here.

[@gumvue.studio]

Hi Tim, I add a full documentary short 8 min film.. film about the Bengal Tiger very similar to what we watch on Discovery Channel :)

[@nimab854]

How can you upscale the images from Flux? They are pretty low resolution and magnetic tends to destroy the quality of flux generated images

[@_madebydave]

Looks really great but i refuse to use AI that uses credits as payments, IMO is the use of credits predatorial af, there is a reason why this is used in so many scammy Gatcha and mobile games. There is no reason not to say you can get so many images for so much $$$, the fact that they don't makes them sus in my eyes. (and looking at their prizing page it is sus af) (yes if you are wondering it is hard to navigate AI tools that don't use this predatorial system, sadly) Great video as always 👍🏻

[@TheoreticallyMedia]

Thanks so much! Yeah, there seems to be a lot of comments on the credit system. I’ll let them know that seems to be an issue. Question: do you think Midjourney’s “hours” approach works better? While I know we’d all love a monthly unlimited plan, I don’t think anyone offers that. Well, besides Runway, and that price point does have its critics as well.

[@_madebydave]

@@TheoreticallyMedia it is the better of the two evils 🤭 I just don't understand why it is hard to go to generate image and then have a calculated amount of money what the image is going to cost (and I am not talking about Open Art AI platform in particular, I am talking in general as I do understand pretty well why they do it in games). I find credit systems incredibly customer unfriendly even more so if you are obligated to get a subscription with no rollovers). I am sorry I just have game PTSD 🤣🤣

[@anony88]

Is midjourney still the image king for realism? I heard the "office hours" update for midjourney yesterday and the guy said that Midjourney 7 is coming soon which is pretty much a "total re-write" and new model for Midjourney.

[@hqcart1]

you cant train the model with the trial credits in the promo code :(

[@coco-openart]

Hi @hqcart1, this is Coco from OpenArt. Sorry about it. Due to the high cost of training models, we don't offer model training with free trials, but you can use the lowest tier to experience it :) I would recommend giving it a try. Thanks!

[@Aiadmiral]

I just jump into and it says "0 day left in your trial, 120 trial credits left" Some kind of bug? I still have no used anything. Just came in, register, jump into discord... and when I have return - boom!

[@TheoreticallyMedia]

I'll check with them on that. I know they're circling the comments! Sounds like a bug? I'll pin this one as an issue though!

[@LouisGedo]

2:01 🤔 "standard"? I'm only aware of 3 versions: Pro, Dev, & Schnell. Decipher for me please!

[@TheoreticallyMedia]

That was my bad- it’s Schnell, I believe. Kinda brain farted there. Oddly, using the other models gave me a bit of appreciation for the older SD models. Still some great outputs there.

[@DarioToledo]

Thanks for the credits but one-time credits platforms with no auto renewal on free plans are a big no-no to me.

[@LouisGedo]

👋 hi

[@TheoreticallyMedia]

Heya Louis! Have you played with Comfy at all? This might be a cool platform for you to dip your toe in, if not!

[@LouisGedo]

@TheoreticallyMedia No......I stopped using local install (open source) when SD 1 or XD (can't remember which) was initially released because I was getting much crappier results than the previous SD version. It was around that time that I got heavily into Midjourney ( even though I'd been using MJ for a wh8le already). Then when Dall-E 3 was released via MS, that was my go-to platform. But now that Flux is available via Comfy, I definitely want to get back into open source local install to take advantage of everything that this convergence of AI model, fork (Comfy UI), freedom (lack of NSFW censorship), and technology (I have a powerful system) has to offer over highly censored and expensive AI tools.

[@LouisGedo]

@TheoreticallyMedia Can you recommend the best way to install Comy UI so that future updates and additions of community LORA's, and other add-ons, etc is simplified ? I have 3 or 4 SD forks installed on my hefty system (which I haven't generally used outside of AI image creation locally) including AUTOMATIC 1111, Invoke, and a couple others that are probably nearly obsolete by now.

[@LouisGedo]

@@TheoreticallyMedia 🤔 I think my first reply to you here was eaten by the insane YT algorithm 😔

[@TheoreticallyMedia]

@@LouisGedoso, I think the best (and easiest) way is via Pinokio. 1 shot install, so it’s dead easy. I think it’s Pinokio (dot) computer

[@antiplouc]

If you cannot run it locally, it's not an alternative. Ai requires you to be able to experiment a lot and create a large amount of images. Impossible to do so with a paid plan that limits the amount of credits you have.

[@TheoreticallyMedia]

Well, that’s what I mean about being a great path from beginner to pro. You can start generating with the various model and “graduate” up to Comfy, and even run that on a cloud server to learn it. When you’ve got the ropes, you can start using it locally, and then share your workflows with the community you learned from. It’s a great system.

[@Ai-Art-42]

interesting

[@aliaslordcaine]

Hi Tim, Thanks for all of the help choosing AI video tools. Sometimes, as you know, the results you get can be hilarious. I thought I'd share this one. It was so funny I made a quick 6 second video out of it: kzbin.info/www/bejne/bX3XgXtnhst3g9E It also cracks me up that I used Hailouai/Minimax, Runway ML, Pixverse to get all of these amazing Zombie clips and then when I went to get simple shot of a Mother and Father giving their son some chocolate ice cream I could not get the shot... maybe I should try vanilla.

[@ECHSBACHS]

: )

[@TheoreticallyMedia]

Hey man! Good to see you!

[@evacuatedspace69]

Great, but you wanted to know what we'd like to see??? A better free plan! 😩😩😞

[@TheoreticallyMedia]

Haha, I know, I know. Part of my new mission when working with companies is to try and get you guys as much as I can with bonus trial or free stuff. I get that they need to turn a profit, but I also totally get that you all are subbed to like 40 services. As am I!

[@dz.audiovisual]

output quality overall seems very weak, lots of wrong things on images in my experience

[@TheoreticallyMedia]

I think once Flux Dev drops, you’ll see a pretty big leap. Some of those older SD models do need some editing, but I guess that’s why there’s that suite of editing tools. That said: flux dev. When that drops, this’ll be a powerhouse

[@coco-openart]

Hi @dz.audiovisual, thanks for your feedback. This is Coco from OpenArt. We're constantly improving. As Tim mentioned, Flux Dev is coming real soon (in 1-2 weeks) and that could make a huge difference. Meanwhile, you can use prompt assistant to improve a prompt to enhance the quality too.

[@pranjal9830]

I don't want to say this but you know most people going to use fake account if they can get free credits to run this...

[@moomoo8115]

I am using this tool for months now. NEW it's not.

[@TheoreticallyMedia]

Haha, new to most of the channel viewers. Although, it’s funny, that almost shows how fast everything is moving. In the span of a few months, something goes from “new” to “old” A month in AI is basically a year.

[@moomoo8115]

@@TheoreticallyMedia As always, it's nice to watch you.. and learn more. Thanks

[@MustWorkWeekends]

Lol all skill levels?

[@TheoreticallyMedia]

Yes sir. Jump into one of those Comfy UI workflows-- that's a skill.

[@AscendantStoic]

Tell me you have never touched or created a Comfy UI workflow without telling me you have never touched or created a Comfy UI workflow in your life. Go back to your cave man.

[@MabelYolanda-c9i]

Why pay for something that is q00% open source and free to use elsewhere? I’d only consider to subscribe if they add video AI and a robust and professional lip syncing, with 16:9 output and key frames. Otherwise is more of the same……and Olivia has to understand that Comfy is DOA because clearly people has spoken and said no to nodes….supporting node based apps is a path to doom……learn grasshoppers from Steve Jobs: “Apple will never adopt node system, period”…..

[@TheoreticallyMedia]

I mean, it depends. Open Source is obviously where so much development happens-- and is super important to the space. BUT, not all users have access to machines that can run all these models-- or, some of them don't want to get bogged down in all of the technical details with installing. They just want to create. On the Comfy side: Yeah, it didn't hit mass adoption for users, but it is still in the backend running a lot of the current platforms. A lot of those AI Upscale platforms are basically just Comfy-- they just don't say it's comfy.

[@KanedaJones]

listening to you lament on what you lost in an image after going further down the AI editing rabbit hole just reminds me how uncritical and accepting people are of things these days. ffs

[@TheoreticallyMedia]

But to note, I mention that we can quickly bring it back with ease. And often times, I actually find that occasionally the “off” results can lead to some interesting places, if you’re open to it. At the end of the day, it’s a new technology. It needs to be thought of more as Photoshop 1.0 than today’s version. It’ll be weird and messy, and purists will complain that it isn’t being processed in a dark room, but all of this will only get better. And artists will find new and interesting ways to utilize it.

[@zddnooig]

cool