Show HN: I used OpenAI's new image API for a personalized coloring book service | Hacker News Reader

Show HN: I used OpenAI's new image API for a personalized coloring book service

clevercoloringbook.com

316 pointsdarajava188 comments

I've had an idea for a long time to generate a cute coloring book based on family photos, send it to a printing service, and then deliver it to people.

Last month, when OpenAI's Sora was released for public use I (foolishly) thought I'd manually drag-and-drop each order’s photos into Sora's UI and copy the resulting images back into my system. This took way too much time (about an hour for each of the few books I made and tested with family and friends). It clearly wasn't possible to release this version because I’d be losing a huge amount of time on every order. So instead, I decided I'd finish off the project as best I could, put it "on ice," and wait for the API release.

The API is now released (quicker than I thought it'd be, too!) and I integrated it last night. I'd love your feedback on any and all aspects.

The market is mostly family-based, but from my testing of the physical book I've found that both adults and kids enjoy coloring them in (it's surprisingly cathartic and creative). If you would like to order one you can get 10% off by tapping the total price line item five times.

Open articleView on HN

Every example is the Studio Ghibli style. Tasteless and no respect for the studio and its legacy.

I don’t understand how you can do this and not feel horrible about it. But I guess not everyone cares as long as it might earn you a few dollars…

I suggest that they use the Disney style. It’s similarly popular and well known but no one will stand up to protect it since it’s a multi billion dollar company that we already know has no respect for artists or artistic legacy.

100% agree, what the AI “techbros” have been doing has been distasteful and downright disrespectful.

Sam Altman is also a little bitch for taunting people like that through his business ventures. First that Her actress’ voice and now this.

Style has been copied since time immemorial and was never copyrightable. I think it's better to be the studio everyone copies in their AI slop than the studio nobody copies in terms of publicity, sales and cultural impact. And the mass copying by a handful of models in a reasonably consistent way together with everyone talking about it is probably better for the studio than flesh and blood artists here and there copying it not so consistently / competently without the buzz attributing the style to the studio

> I don’t understand how you can do this and not feel horrible about it.

I don't understand why you think one should feel horrible about generating images in some visual styles. What's the problem?

Demonstrably, it's not something that's generally considered protected - it's not in the laws, and I've got this impression that the request of "$artist/$studio-style art" was generally considered socially acceptable. AFAIK it's also a part of academic courses, where artists practice various styles.

Patron requests, homages, pastiches - all this stuff existed for a long while and was generally accepted (or so I think), the only difference is that a machine does it now, incredibly fast and cheap. People used to hire artists for this kind of stuff since times immemorial. Nowadays, if a machine can do a passable job, then why waste human's most valuable resource (time) for it.

It would be interesting is to hear Studio Ghibli's opinion on the matter. Not someone who thinks they might be wronged somehow (no offense meant, I do not intend to invalidate your opinion) or someone who rather thinks they might be even benefiting from this - I'm sure it's likely to be a multiple-edged sword, as life is rarely simple - but their own actual thoughts on the subject. I wonder if they already published something...

Good news - we have commentary on that. Hayao Miyazaki calls artificial intelligence animation "An Insult To Life Itself" [1]. It's one of his more well-known quotes.

1. https://www.indiewire.com/features/general/hayao-miyazaki-ar...

Thank you.

I have watched the video but, sadly, it seems to be on a merely tangentially related subject (some zombie ragdoll movement), that caused an understandable disgust (as explained in the video). Then there was a comment about the desire to create a machine that can draw pictures - and I’m really not sure how that’s connected to anything that happened in the video.

The only thing I was able to extract from it is general disapproval, but that’s about all I can be sure of. Better than nothing but not something I was hoping for.

Upd: oh, I haven’t noticed this immediately, but it’s a 9-year old video. I was hoping for something more up-to-date, as a lot of things had happened since then.

What's so bad about it?

"Imitation is the sincerest form of flattery that mediocrity can pay to greatness."

“Whoever creates this stuff has no idea what pain is whatsoever. I am utterly disgusted. If you really want to make creepy stuff, you can go ahead and do it. I would never wish to incorporate this technology into my work at all. I strongly feel that this is an insult to life itself”

Though I agree with the sentiment that Miyazaki must hate OpenAI commodifying his style, I find it so ironic that people misuse a quote of his from 2016 to be about AI generally. Another kick in the nuts for the poor guy.

He was talking about how an AI generated gruesome animation made him think of his friend awfully struggling with a disability. That people who create that kind of stuff don't understand pain and suffering.

The full quote: "Every morning, not in recent days, I see my friend who has a disability,” he said. “It’s so hard for him just to do a high five; his arm with stiff muscle can’t reach out to my hand. Now, thinking of him, I can’t watch this stuff and find it interesting.

Whoever creates this stuff has no idea what pain is whatsoever. I am utterly disgusted. If you really want to make creepy stuff, you can go ahead and do it. I would never wish to incorporate this technology into my work at all. I strongly feel that this is an insult to life itself”

Most people never heard of Studio Ghibli and their style. They just like how it looks. Whey you have only one such book at home it's not annoying.

As for the project: it's nice and easy, likely will be replicated soon.

This is a cute and simple idea!

I'd like to see what a real physical book looks like before I buy it though. Do you have real pictures of a printed one?

I think our kids would appreciate seeing the original (even if a small thumbnail) along side it. You can't always tell from these AI drawings that it was originally you and your family.

Also, it's REALLY expensive. $30 for a book that my kids will draw on in one or two nights and then never touch again is probably too much.

Thanks! I've added a section at the bottom of the site showing some real photos of an actual coloring book I got in the mail. There are thumbnails of all photos uploaded on the back.

$24 + postage is the lowest I could reasonably charge for this. Printing costs are a bit more than half of that, OpenAI charge a surprising amount for image generation, but there is also a good amount of human effort (and creative choices) in generating the book. It's not a fully automated process and I hope that's evident from the quality of the end product.

It’s not cheap, but my kids treasure coloring books for a long time and probably one like this until it falls apart.

To the author, I have this idea, for each page, put a sheet of transparent plastic or something like that. So the owner will color the plastic which can be erased. But it may increase the cost anc the color may not stick to the plastic.

The comics look pretty Miyazaki-inspired, like all of the comics I've seen lately. I've kinda started to dislike this look because it's _everywhere_ that low-effort comics are these days.

Maybe worth trying to train a better style for this. This is probably something where you could put a little effort in up-front (ie: using a model that's for segmentation to get outlines, using some classic image-processing for boundary detection) and then have AI touch it up a little more lightly and a less of the "default" style.

Also, do you have AI images for the "real world" samples on the left? They have a certain "I don't exactly know what, but it's creeping me out" vibe.

It doesn't look particularly Miyazaki style to me; it's just a generic cartoon style.

I think the Ghiblipocalypse has gotten people on edge.

OP confirmed that their prompt includes a directive for Ghiblification. Given that Miyazaki is known to hate GenAI I really can't condone... I mean there's nothing anyone can do about it but it's just kind of sad.

>Given that Miyazaki is known to hate GenAI I really can't condone...

Miyazaki also said (in response to the Charlie Hebdo murders): "For me, I think it's a mistake to make caricatures of what different cultures worship. It's a good idea to stop doing that." I still love the man and his work, but he's not some infallible authority on what is and isn't appropriate in art.

Seems like a perfectly reasonable value statement. I am inferring that you see that as a call to censorship, but unless there's some additional context here that suggests that, I don't think that's the case.

> Given that Miyazaki is known to hate GenAI I really can't condone

Not that I wouldn't similarly expect it from Miyazaki in terms of general generative art but the actual source of all the articles/memes about his quote point to a 2016 video where he's being demo'd a disturbing 3D simulation of an oily looking human figure crawling on the ground by its head while the dev explains to Miyazaki and others that 'it feels no pain so it learned to move by its head' and it could be used for horror games.

It's then that Miyazaki expresses the 'insult to life itself' quote and explains the devs have no idea what human pain is. Makes one wonder how the devs thought the reaction would be any different tbh.

Edit: reading that he clarified in an interview[1] a couple years later that his distaste was due to believing the dev was aiming at humorizing such body contortions of realistic humans which he took issue with.

[1] https://realsound.jp/tech/2018/10/post-270755.html

Hm! Thanks for the fact check. I really do want to know what Miyazaki actually thinks of the Ghiblification thing at present.

Yeah can we stop spreading this misinformation. His quote was in reference to grotesque nature of something he saw.

> Miyazaki was shown an AI-generated character. The character was a scary monster that used its head as a leg because it couldn’t feel pain. The person presenting it said its movements could be used in making a zombie video game.

To which he stated:

> Every morning, not in recent days, I see my friend who has a disability. It’s so hard for him just to do a high five; his arm with stiff muscle can’t reach out to my hand. Now, thinking of him, I can’t watch this stuff and find it interesting. Whoever creates this stuff has no idea what pain is. I’m utterly disgusted. If you really want to make creepy stuff, you can go ahead and do it. I would never wish to incorporate this technology into my work at all. I strongly feel that this is an insult to life itself.

The author has since listed the prompt elsewhere in the comments. It includes, "The drawing is in a simple Studio Ghibli portrait style."

https://news.ycombinator.com/item?id=43801189

there's 4 sample pages, and the one with the cat is the only one that is not Ghibli-style.

Here's some generic cartoon styles to look at: https://i.pinimg.com/originals/5f/04/ef/5f04ef77ce3beb272a61...

That's probably the only info-graphic I've ever seen that's worth a damn.

This is so interesting - I could add some of these as options.

Jesus Christ

It absolutely resembles the current Miyazaka-esque OpenAI image trend that’s been going on.

This has zero resemblance to Miyazaki’s style. (And I say that as someone who isn’t a fan of this idea at all.)

Hard disagree. Sample #2 is totally the Miyazaki-vibe that is everywhere in OpenAI-generated comics. https://clevercoloringbook.com/samples/2_cartoon.png

The cartoon owl at the top has a different vibe and would probably work for the comics as well.

After seeing this example, I think this is the elevator pitch: ”We take your personal highlights and make them as generic and impersonal as possible.”

It's nothing like Ghibli, you are overthinking this.

You think that the model is not properly producing the style it was prompted for in the below prompt? I don't see other artists or styles mentioned.

> Make this a page in a colouring book. The drawing is in a simple Studio Ghibli portrait style. Bleed all the way to the edges. Background colour is #ffffff and lines are bold and #000000. There is no shading or crossthatching.

I'm gonna agree with the above comment - #2 looks like a Japanese-style cartoon (for better or worse).

OP confirmed their prompt explicitly asks for a studio ghibli style

from clevercoloringbook.com:

    > Please only upload photos that are in line with OpenAI's Usage Policy.
    > We are not able to include any photos that do not follow their policy in the final printed book.

from openai.com/policies

    > Editing uploaded images or videos that contain real people under the age of 18 is not permitted.

The first two sample pictures on the page contain of adolescent children. Are you concerned about this apparent contradiction?

Great point. As per our TOS - users of the site must be over 18 and have the consent of everyone in the image (i.e. their own kids, relations etc).

I put that line about OpenAI's usage policy there for practical reasons. If someone orders something that OpenAI refuses to generate (like a photo of Bart Simpson say), then I can't include it in the printed book. With this project, if someone uploads content that's in any way inappropriate, we'll see it and refuse to fulfill the order (and take other appropriate actions, if needed)

My reading of the policy is that if the real person is above the age of 18 then they can consent to the editing of images which depict themselves under the age of 18. So I could upload and edit a childhood photo of myself because I am a real person above the age of 18 and I am the person "contained" in the image. But I couldn't do the same for my niece who is a real person under the age of 18.

I'm not the OP, but during the recent Studio Ghiblification craze there were a huge number of photos of families and kids passing along in facebook, twitter, and other social media. It was literally everywhere you looked. OpenAI obviously saw all of that. I don't think they actually care unless it's something bordering on illegal.

I agree. In practice OpenAI is unlikely to care about families uploading their own photos. I think the policy is mostly to stop random people from engaging in creepy activities with the photos of children.

> "that contain real people"

It seems the loophole on this site, is the examples (by my best guess) are AI.

People are complaining about using the Ghibli style, but I'm here to complain about the fact that they turn everything into generic abstractions that look nothing like the people in the original photos. Generic-cartoon-vaguely-reminiscent-of-your-family-but-not-really is a product that I'm surprised someone would be proud of.

I wonder what would happen if you ghiblified images of the slop I cooked for dinner. Ghibli food always looks so good

For anyone looking for a prompt to do this manually, it seems to be as simple as this:

> Generate a version of this photo that can be used as a coloring sheet

Close enough! The prompt I use is:

I think one thing that slightly drags this down is the Ghibli style produces images that don’t really look that much like the original people. If you can find a prompt that is still stylized while preserving more of the characteristics of faces would go a long way on the personalization front. Maybe easier said than done.

Anyone else feeling this weird vibe in the age of AI? You get a cool new idea, but then you think about how easy it is for someone else to replicate it, and you end up not doing it.

Often multiple times a day, but on top of that concern, also throw in various ethical concerns and a few questions about if it actually has value or should exist

likewise, at this point I am just giving in and making what I can to see whether or not it is feasible, to speak nothing to its value or existence

Me too except I usually don’t get very far into making the things

Yeah I've been doing this with image-gen AIs pretty much since they started and it's a lot of fun. Even early Dall-E etc was awesome at doing stuff like "Create a colouring in sheet with some dinosaurs having a party" or generic prompts like that, and more recently giving photos to convert has been loads of fun for the kids.

The OP looks like it runs images through a Ghibli filter first

I’ve done this a bunch with my son. It’s not quite that simple because often times it’ll create images that have too much detail, sometimes it’ll actually include colors. But yeah, it’s not really all that complicated

For what it’s worth (and it’s probably not much), it doesn’t cost that much to commission comic book-style art from an actual artist online. When you do that, the proceeds go to an artist, not to an AI company that stole from them and a software developer who wrote a wrapper around their API.

In fairness no artists are advertising a personal coloring book. The time, effort and cost would put this out of reach for 99.99 of people.

No artists are losing income because of this and no industry is being upended. This is a new product that's available because of a technology advanced.

Why the focus the artist? Everytime you order in food online you take away a tip from a host, server, bartender and take away a job from a person who answers a phone. Why focus on artists when so many have been affected by technology.

And yet there’s plenty of adult coloring books made by a human out there if you’re willing to go to a brick and mortar shop. Got a super cool one from dick blicks, with a lot of underwater scenes. Also paper quality is important. I can’t imagine getting as far as I did in mine if it was newspaper

That’s because those are not personalized. The economy of scale allows for artists to make generic coloring book with high quality art, but it’s expensive for artists to create (and customers to buy) custom made coloring books personalized for the customers photos.

My partner makes one! Go grab a copy if you're in Australia, the wonderful POP local -- started as POP Canberra -- sells them.

https://www.poplocal.com.au/product/bum-man-colouring-book/

He's 'Bum Man'. A man (actually it's asexual) who is a bum. I mean c'mon.

Looks cool, but they should show more of the pages online. All I see is the (already coloured) front cover.

The food you order online was not stolen from the server/bartender without their permission or compensation. Even if the analogy holds, this is whataboutism, and in the U.S. at least tipping is a fucked system too.

If you stop going into the restaurant they stop scheduling servers. You or the restaurant didn't get permission from the server who isn't working there anymore.

It's about applying your outrage evenly. Why put artists over a servers? Why do you drive when not using horses means many blacksmiths positions disappear. Technology that is accepted by society changes society. Artists will continue to evolve and create messages about those changes. No need to worry about their plight. Worry about translators or other industries that can't easily provide the same value. Artists are the one group who will survive and thrive.

You’re right that the food itself wasn’t stolen, but how many restaurants actually come up with their own recipes? And how many use recipes created by master chefs that were ‘stolen’ and used by others?

This is how art works and has always worked. Artist should be using the same AI tools that the general public use but create things that the general public cannot. That’s what artists have always done.

Culinary skills at the high-end are typically passed on directly from chefs to their apprentices, intending to be used, built on, and passed down again. It doesn’t really work the way being described, which is fine, because there’s no way to shape this scenario into a comparable one.

Any attempt to compare the A.I. stuff to some analogous scenario is deeply flawed if it does not include 1) that A.I. instances are not humans, but computers run by companies, and 2) the incredible scale at which it can operate.

The actual actions taking place are secondary at best, and the situation cannot be judged on that alone. It must be debated in the context of the actions being undertaken by machines, owned by companies, motivated by profit/market share/growth/whatever, with little communication or collaboration with the humans who created the works, and that they can now generate outputs based on those works at a scale, frequency, level of precision several orders of magnitude higher than a human can ever compete with. It cannot be compared to any sort of person-to-person scenario. The enormous scale this operates at, by actors that are not human, is the core of the situation.

Recipes are the least of what goes into a restaurant. It's not a secret. In many restaurants the chef will give you the recipe if you ask nicely. If not, anyone skilled in the art could reproduce it.

Running a restaurant is a trillion other things. Ordering the right amount of ingredients. Hiring, training, and keeping staff. Cleaning the bathrooms. Replacing stolen silverware.

You're not paying for the secret recipe. There isn't one. You're paying for the insane amount of work that goes into putting cooked food on a plate.

Images are much more about the specific process that went into creation. The intellectual part that can be taken is a much higher fraction of the product.

If it's not plant based it is.

I tried to do exactly that once. I was offering between $20-$40 per image to make a few coloring pages as a mother's day gift for my wife. Not complex images either -- just basic coloring pages from photos of my wife and child, without backgrounds, for my kids to color in.

I reached out to multiple artists, and got one image back (from a good friend). I gave up on commissioning actual artists, and traced the images myself on a tablet. I imagine someone with the right knowledge of where to find artists and the willingness to wait on their schedule could have done it faster, but I'd have used this service if it had been around.

If this person’s service was to pay human artists $24 for a 23 page custom coloring book you’d be crying on here about them not paying human artists enough.

Almost nobody is paying $100 or more for a custom 5-page coloring book.

This service isn’t taking work from human artists.

Usually when you commission something you're asking the artist to do art and create something unique with their own artistic flair... not just line-trace an existing photo.

The intention and cost of something like that is not at all comparable to what is being offered here.

I can't imagine how much it would cost to commission an artist to do a whole coloring book and then organize them and send them to print but it's a good point. AI is never going to be as good as a real commissioned artist, but this idea makes having something similar far more accessible to a lot of people.

AI will exceed the capabilities of human beings in any task given enough time.

Sure, not now. Not tomorrow. But less than 1000 years from now? Definitely, imo.

My opinion isn’t fully formed but I currently think either all content producers have a claim (potentially workable as eg a discount), or only those who contribute should get access to AI’s.

And by all I mean the AI companies owe a huge debt to all humans who wrote or designed or drew anything. The vast majority of the benefit of this technology relies on volume: the billions of pages and lines of code we wrote for other humans, but have now been repurposed. This technology relies on bulk, which was mainly unprofessional or freely given content, by those who intended it for other humans. It was not 100% built only on the output of the few who charge for their exquisite words or designs, even if their output is higher quality.

Alternatively, let the AI companies go for it but everyone who uses any kind of AI should understand that they’re standing on the shoulders of the millions of developers and nonprofessional writers whose work has now been repurposed. Not the few artists and journalists. So those artists and journalists should both refuse to contribute to, and use, AI.

* I’ve written very little of this useful content, but would be happy to pay my share to those that have built what we have. I also turn off training on my content, but I pay a lot for models. Feel free to help me think through this with comments of your own.

I am following a similar mental path. I feel like the AI companies should be paying some sort of tariff on their output, going towards everyone on the planet who contributed anything at all. I don't think you can account for it more finely than that.

Yup, it feel similar to mining. The models are a transformation of the underlying content. The content is owned by “us”. A mining company can come in and transform rock into iron but it owes something to the grantors of the mining rights.

There is certainly a contribution in improving how the body of work is represented but we treat the “AI” as the smarts, when really it’s a lens on the collective knowledge we have built. You can make the lens better, bit not claim ownership of the body of work. Right now that’s what’s happening, with a few edge cases for artists and publishers.

If it does not cost that much, that is obviously because the artist is too cheap. If you find that to be a preferable equilibrium, that's a choice I guess, but I find it fairly ironic in light of the purported motivation.

Maybe, but then I have to negotiate with the artist, handle their refusal to draw art of my choosing, and wait for their (possibly unpredictable) schedule. AIs mostly avoid these problems.

Those transactions never would have happened, and never will happen.

I can get ChatGPT to do this for literally free. Even in the free tier, I can get a couple images per day.

Didn’t the artist “steal” from artists that came before them by looking at and taking inspiration from their photos? Especially ones that would do such artistic genres as commercial coloring book art?

Yes. But they are people, perhaps with families to feed. Not computers.

Cool idea. I can see keeping colored pages of these by my kids up on the fridge a lot longer than what’s on there now!

So it’s ok to steal as long as you are feeding a family with it? :) I get what you’re saying, anyway and it’s an important distinction.

I guess what you see as “stealing” I see as inspiration. I also believe that there will be artists who use these new artistic image generation models in ways that are new and interesting just like the first person who used spray paint for graffiti was ripped off by everyone else.

It delighted my kids to see themselves depicted in coloring sheets in situation where they are currently interested. There is no world where I would have paid an artist to make these photos, and we would have just colored on blank paper.

Again, I get that real people’s content was devoured by these big companies, but at the same time I am much more concerned about bigger issues and would rather focus on getting ahead of AI rather than fighting it.

As a gay dude I don't fall for the "but they have a family" thing, it reminds me that I'm worth less in many eyes for that reason.

Anyway, the problem isn't with automation, as per usual it's with animalistic, selfish human beings who would rather hate each other while the rich get richer, rather than vote to implement a well needed framework to set up UBI, proper taxation of corpos, etc.

But nah, that person has a different skin colour than mine, that person has different sexual preferences to mine.

I do hope the machines take the reigns at some point.

> As a gay dude I don't fall for the "but they have a family" thing, it reminds me that I'm worth less in many eyes for that reason.

I don't think you're worthless at all. And families come in many shapes and forms.

Thanks <3

Lmao yes it does. Wtf.

I'm a furry, I commission a decent bit of art and furry artists (unless they're super popular) tend to actually be much cheaper than normies.

Commissioning a comic is >$100 for a page, from a popular artist at least several hundred.

And that's also for personal use with no commercial rights whatsoever - it's actually an impasse because they technically still own the art, you technically own the character in the art, so it's in licensing purgatory, which is fine for conbadges, smut, icons and whatnot.

I used to work at an agency that shared a floor with an art studio, commercial rights for art, especially something as complex as a comic page can easily run into $1-10k or more.

Don't get me wrong, artists deserve the money they get, but as with everything that gets automated, there's a financial incentive to do so. Inb4 someone drops all the etsy links of sellers doing AI art as well or (as has been done for a loooong time) using a non-ml based filter to achieve the same thing.

There are definitely cheap-er options available from Brazil, China, Venezuela etc (same as fursuit commissions) but that's also another interesting topic in relation to ai; we already outsource heavily.

This is a cool technological feat but what is the cost to humanity and its artists?

Some of these replies seem rather dismissive to the artists’ plight.

They're dismissive because we've had the same moral panics before with the introduction of photography, then sound recordings, and then digital art tools, and then vector art, and then 3D, and also the Internet to an extent, and...

You can see where this is going, right? In the end, humanity and even artists will be fine overall, even if the world changes.

Nah it'll be the end. And welcomed, because humans are too apathetic to vote for real issues like UBI.

We'd rather be tribal and hate on race, gender, sexuality, religion, location, etc than actually act our neuron counts.

This feels different to me. This isn’t the camera going from film to digital. This is the camera taking over the photo creation process, and developing them, and selling them. What’s the point of the human?

It feels different because you are living through the change. What happened before is something you read about and already know the outcome. Future generations will say the same.

how will artists be fine when Google can steal all their work, then use that to compete with them and ultimately replace them

for the cost of showing ads?

Cost is nothing because this service isn't offered currently. No income lost and might spark an interest in coloring books which grows the artist's income.

Artists have been around and existed in more repressive societies throughout time. The best art is usually produced from the greatest struggle. Artists will engage and create art in this new world. The cost of not providing a new surface for artists to explore is what kills art.

How does buying an AI generated coloring book lead to more sales for the artist that doesn’t offer AI generated coloring books?

“learn to code”

Just code the food!

This is fucking amazing!

Everyone and their mother are trying to hop on the band wagon of AI and make a half assed service just because it may sell just due to the "ai" tag attached to it - this is different!

Chapeau bas! It's simple but brilliant. It's a great example of what a good idea is - with minimal effort he made an epic product focusing not an AI, but what AI can bring to the table and executing it flawlessly. Hats off!

i've been doing this with my child and picture books. i take pictures of pages, convert it to a coloring book, print it, and then we color her favorite books together.

You're uploading your family pictures.. nevermind. Go enjoy your coloring book.

Thanks so much! Really glad you like it.

Why can’t people just appreciate some thing cool rather than turn every HN comment section into a controversy? I had never heard of this Ghibli controversy.

When you learn of a situation that a lot of people care about and feel very strongly about, it would be wiser and more respectable to not immediately dismiss it. Since you admit to only finding out about the situation five minutes ago, you might not have a firm enough understanding of the nuance to reduce everyone involved to haters without valid concerns. If you dig into the subject a little more, the answer to your question will become apparent, even if you disagree it.

I don’t agree or disagree with it. It’s the constant outrage. I don’t understand how people can be constantly unhappy

You don’t know if anyone upset about this is a generally and constantly unhappy person, they are upset about this. Are there no topics in this world that you have a very strong opinion on, that you would express given the situation? Could you not be a generally happy person and express those opinions, at the same time?

Also, I mean, I know a lot of people don’t care about the whole “A.I. vs. artists” thing, but it should absolutely not be difficult to understand why many do. We are talking about a fast-growing technology and industry that will perhaps decimate jobs and entire professions, that will definitely reduce the value of certain things to zero, and while that will be good for some things, it is concerning for many that one of the first things being seriously threatened is art—something generally thought to be a deeply human ability, and a profession already notoriously difficult to earn a living at. For a lot of people this is existential. This guy’s little coloring book project is not the problem, but it‘s still a small facet of the larger issue, and being concerned about that issue is very valid, and anyone with perspective and a modicum of empathy should be able to understand it.

Thanks for the perspective. Makes sense

Maybe I am thinking a bit meta here. But who is supposed to colorize these pages? Kids that are in the progress of learning to instead use AI for everything? It can surely deliver better results quicker, after all.

Same question would be relevant if you wouldn't have used AI to generate these outlines, of course.

I just want to point out there is a certain irony of the "cut the branch you are sitting on"-kind here.

Edit: typo

Congrats!

I would like to know the cost of the tokens you are paying for an image. How many pages coloring book will be created against $24 book?

Thank you!

OpenAI costs are surprisingly expensive. It's about $7 to generate a whole book (24 pages). There are 8-24 images allowed in a book, with a cover too. So there'll be 48 max pages in a book (incl blank pages).

Thanks for your prompt and kind response. Yeah, it is expensive because it is new at the moment. They already mentioned the high token cost.

Oh, sorry - I'm going to be that person....

Your step 2 is wrong :-) > Step 2: We convert them into a high-quality physical coloring book with OpenAI’s brand-new Sora model, then send it out for printing.

You don't convert it into a physical book /before/ sending it for printing.

Oh, lol. Nice catch - I'll fix that.

Not sure if you're aware, but if you're interested in SEO/AEO marketing, there's very healthy monthly traffic for long-tail searches in this area. Some searches getting towards 100K per month.

Love the idea! Good luck.

I don't know much about it, feel free to send an email to dara@clevercoloringbook.com

Where do you get that from?

You'll want to really drive home the niche (through your feature set) that it's for family photos, because the generic photo to AI vectorized coloring book service has been done to death.

Can you show me an example? I think before 2-3 days ago this wasn't really possible without the output images looking really bad.

I've been using (mostly) the OpenAI image generator for quite some time generating coloring pages, it's pretty decent at it and can generate just about anything my kids want as long as you word it a bit neutral to avoid it generating (something it recognizes as) copyrighted content.

Great idea to turn your own photos into a coloring book generator!

Edit: I wonder how you prevent it from generating copyrighted content when people upload e.g. 'photos' of Disney content? Or has that not been a problem yet?

Thanks! Hasn't been a problem yet, but if that happens I'd just email the customer directly and ask for a replacement photo/if it's okay to leave that one out.

Cool idea and really nice looking site.

Pricing is quite high - 24 pages maximum for $23.99. There are 100-page coloring books on Amazon for $5.00, and the age group that really would be using this is not going to remember what was on the page a week from the day they did it.

Maybe it can work in the nice of "adult coloring books" - I've seen some social media content where people really go crazy on coloring books, and being able to get nice physical copy to work off could appeal there.

Presumably you aren’t the parent of a 5-7 year-old child. I might try this manually and save some money but my kids will absolutely cherish coloring themselves, their friends, and their parents. We’re on vacation now and this is gonna be big when we get back.

Interesting presumption. I know about the 100-page coloring books because I've bought them. Paying $1 per page at the speed they get colored would cause me to go bankrupt. I presume you're fabulously rich, and it doesn't matter.

I stand corrected. I think we are looking at this product through different lenses. E.g a page from this coloring book might wind up framed and kept forever, whereas all our other coloring books are eventually secreted into the trash.

You may want to give dibulo.com a try. We also launched a photo -> coloring page, but we will go one step further that you can bring it to life again.

My 5 year old will start coloring, get bored and scribble all over it.

Thanks! I priced it as low as I could given the costs of printing the book, Sora's API costs, and the human effort that goes into it (there are some creative choices to be made too!). The 100 page books you see on Amazon aren't personalized and probably not the best quality. I'm also hoping a completed page from one of these books will be a nice keepsake for the parents as well as being more of an incentive for a child to exercise some creativity.

Love it! Idea: option to print "mini books".

I have some kids that still color, and it would be great to keep something in my pocket to give them quick with a crayon or pen.

I was thinking of doing different sizes alright! They might not be too much cheaper, that's the only problem I can think of there.

I'd pay for the convenience. Not sure about others!

Suggestions:

* one full PDF (including cover) of an example book.

* don't use AI images as examples - it's not obvious if the outline version will look as good on real images.

Thank you, I agree and I'll update the example photos.

I didn't add a PDF but I added some photos of the real end product to the bottom of the landing page now.

I like GPT wrapper's that let me personalize/customize existing real world things, and this a good example of that. I like it.

Thank you!

I'm from India. I was looking for a service like this but without sending me a physical book part.

I'd use this at @ $10 price point if I'm able get downloadable a4 coloring pages from a picture. It would be great. Also this way your customer base becomes international.

$10 for 20 pictures is a good price point for me. Pretty expensive but I'd still go for it.

I'll implement this today.

Edit: I've implemented this! It's a lot of code changes so I hope I didn't break anything.

Thanks!

This is an awesome. I had a similar one: Convert the dense non-fiction books into something more readable. eg. SAPIENS vs UNSTOPPABLE US.

But this makes me wonder: What is the barrier to entry for these apps now? Anyone can do it. There is going to be a barrage of apps/websites like this?

Why would you do that to Sapiens, it's an enjoyable read

Seems like this cat (and various variants in similar settings) was a top rated image in Sora's explore/images a week ago. Was it yours, should it be credited, or did you hit edit prompt<enter> to get a variant?

No worries, just wondering how that should work.

Wow, yes you're right. I did in fact take that image from the Sora homepage because I thought it was cute.

I pretty much just assumed they're all in the public domain. I can't find the image anymore so I've decided to remove it for now. I generated the other three myself.

They look generated, which is off-putting. Maybe use three real photos instead?

How did the launch go? Could you share Revenue stats after the first 24 hours of the launch?

I tried copying the one of the photos into ChatGPT and asked it to make me a coloring book style image out of it. It is definitely not as good as this site (i.e. I don't think kids would recognize themselves). Good prompting!

Ask GPT to give you a prompt that would create a color book :-)

I always use LLMs for meta-prompting. They know themselves better than others :>

Nice and simple! I'm excited for all the fun micro businesses that get enabled by the new image API.

Things like your coloring book, instant sticker/tshirt/swag creation, video game assets, etc.

Also love the "tap 5 times for a discount" feature.

Thank you! I was happy with the discount thing alright but it's not useful for tracking referrals unfortunately.

Wow a lot of criticism. I'm considering a similar business. I think this is too expensive when printing this is so easy these days. But charging some small about per printable coloring book would be very attractive.

The printing aspect of this wasn't too easy... I wish I could charge less but as it stands (especially with surprising API costs) I'm barely making a profit on this.

have something similar in the building pipeline, yeah the API costs caught me off-guard too. I knew it was going to be expensive but this is still pretty high

Well done - I like the style of the page and simplicity!

We recently created one too, where you get a printable version: dibulo.com/editor - the next step will be to bring the templates to life again.

Love it! Finally a product that properly leverages AI :)

Thanks!

This is a great idea and looks well done.

I wonder if printing services (Lulu?) have a automatic API or if it requires some manual intervention? (And the shipping part?)

Thanks! There is a good bit of manual effort involved on our side - we generate the images, regen any that don't look great, choose the best cover, and then send the PDF to Lulu. It's dropshipped, we never see the physical book.

Nicely done, Ive always wanted something like this for my family pics. you think AI-generated art will ever feel as special as something handmade?

I mean it's pretty close. The handmade element comes from coloring it in.

Everything is a copy of a copy of a copy of…

Awesome idea, implementation and design!

Thank you!

They have an api now? amazing! I have been using gemini flash but results are... less than stellar.

Yes! https://openai.com/index/image-generation-api/

I was very excited when it came out. Google have Imagen 3 (is that the same as Gemini Flash?), but you need special access to be able to edit images. I haven't tested it yet but I think it's a lot cheaper than OpenAI

imagen 3 is a standard diffusion model the new one is attached to gemini 2.0 flash experimental. model naming and working is super confusing there

Gemini has released new models, maybe they are better.

Also, I request you to expand further, why Gemini is not better?

instruction following is good most of the time, but hands are still regularly coming out garbled and there's a small percentage of failed generation in the mix. style following is also a bit hit and miss, I'm producing watecolor mostly, but every now and then a generation is photographic

site is not really ready for this, but if you want to get an idea go here https://d2uua1ig5ocnwt.cloudfront.net/ and search these:

amber ladybugs north

I manually remove failed gens, but the proces doens't scale to where I'd like if I have to manually verify images. and I have to have a few automatic rejection in the pipeline, like top border not white or image not squared.

Do you ever handle the physical book, or is this a fully automated drop-shipping operation?

There is a good bit of manual effort involved on our side before we send the book to print - we generate the images, regen any that don't look great, choose the best cover, and then send the PDF to Lulu. Yes it's dropshipped, we never see the physical book.

Can you preview the images? I uploaded one image and don't see the outline-version.

No, it's not a real-time process. We have an admin panel to generate the images, cover and PDF. There is a good bit of human input involved.

Why don't you use canny/HED filter :O? (seems pretty overkill for this job..)

Do you not think the AI output looks far more polished and print-ready? Canny edges have a lot of noise and don't look at all clean for coloring book purposes.

Canny filter: https://i.imgur.com/Ju7Ggac.png

Sora: https://clevercoloringbook.com/samples/3_cartoon.png

I've had this idea years ago and searched extensively for a way to turn images into nice line art, but it turns out there needs to be a good bit of creativity (AI) to do so. Old school computer science techniques don't cut it.

Your contact form is not working . Try using FabForm !

Oh nice catch! Fixed, thanks.

Is there somewhere to download a PDF to print out?

No, but perhaps I could add this as a paid option.

great idea! how did you realize the actual producing of the book? you connect to some print on demand api like from printify?

Yes I use the Lulu API. It's great!

Why not just an option to print the image?

You can simply open up Chatgpt and generate the image yourself, faster than it’d take to transact with this third party. The cool thing is that they are printing a physical book for you.

Presumably, it’s more profitable to sell a physical book, which makes sense.

I've added this option - we send out two print-ready PDFs for the cover and the coloring sheets.

For those interested in building something similar, I prompted a story book generator using v0 and Gemini’s image generation a few weeks ago:

Demo: https://v0-story-maker.vercel.app/

The chat: https://v0.dev/chat/ai-story-book-creator-zw7TrmkN2Eb

I love the method you used to get image consistency: construct a conversation history with the model where it explicitly accepts each image and agrees to use it:

  User: Previous page prompt: [Previous Prompt 1]
  User: [Image 1]
  Assistant: I've received this image and will maintain visual consistency with it.
  User: Previous page prompt: [Previous Prompt 2]
  User: [Image 2]
  Assistant: I've received this image and will maintain visual consistency with it.
  User: Previous page prompt: [Previous Prompt 3]
  User: [Image 3]
  Assistant: I've received this image and will maintain visual consistency with it.
  User: Please maintain character appearance, art style, and color palette consistency with these previous illustrations when creating the next image.
  Assistant: I'll ensure the characters, art style, and colors remain consistent with the previous illustrations.
  User: Generate an illustration for a children's ABC story: [Current Prompt]. Make it colorful, child-friendly, and in a consistent style with any previous images. Include diverse characters with different ethnicities, genders, and abilities. Ensure representation is natural and authentic.

I'm curious: did you settle on this after trial and error? I mean, did having the assistant explicitly agree increase consistency, compared with just asking nicely at the end?

I didn’t write that - v0 did. You can see it all in the chat

I noticed it wasn’t passing the image back and forth so I asked it to, and it wrote that prompting

BTW - thanks for sharing the chat. I'm going to study it for prompting tips. It's good to learn from other people's iterations using coding agents, and not just my own.

The quality of the images is incredible! Great job

How much does the api cost to run? Do you have any safe guards in place in case bots try to build 1000 stories?

That’s all Gemini, I did very little prompting

I’m just using Geminis rate limits, because its pretty ridiculously cheap. You can get pretty far on just their free tier last I checked (when I made this the image model was 100% free)

I wrote about it a little bit here: https://x.com/max_leiter/status/1906492622551884187

wow!!!!

Any idea how much it costs to create a book?

Reply on news.ycombinator.com