RoughCoronet0
Dragon Lover
Here is the final set for my Corrupt Demonic Rainbow Crystalline Dragon end boss pics. I added anime to the prompt, and got much more vibrant pics from it. Now I'm going to bed. 
D&D General DALL·E 3 does amazing D&D art
- Thread starter M.T. Black
- Start date
pukunui
Legend
Trying to make "Jane Vanilla", a human evoker wizard and the next Mad Mage PC. No luck with the full length view while including a reference to something in the foreground.
Prompt: Gritty comic book art style, full length view, tall thin bookish powerful female human woman mage wearing dirty grey robes. She has a magic wand and leather-bound spellbook. She has magic nightvision goggles on her forehead. She has magical leather gloves. Her hair is black but she has dyed it red. She is reading her spellbook as she wanders down a dark dungeon corridor with a tiny globe of light hovering by her shoulder. A rat scurries away in the foreground.
Prompt: Gritty comic book art style, full length view, tall thin bookish powerful female human woman mage wearing dirty grey robes. She has a magic wand and leather-bound spellbook. She has magic nightvision goggles on her forehead. She has magical leather gloves. Her hair is black but she has dyed it red. She is reading her spellbook as she wanders down a dark dungeon corridor with a tiny globe of light hovering by her shoulder. A rat scurries away in the foreground.
SakanaSensei
Adventurer
pukunui
Legend
Boom!
Prompt: Gritty comic book art style, full length view, thin scholarly female human woman mage wearing dirty grey robes. She has dyed red hair, and she has magic nightvision goggles strapped to her forehead. She wields a magic wand and reads through her magic spellbook as she strolls down a dark dungeon corridor, her shoes crunching on debris. In the foreground, a rat runs away from her.
The (US-only, extremely topical cases) solution of having AI art not protected as IP would solve this problem since the plaintiff couldn't claim anything by lack of human autorship. In other jurisdictions, there is an exception to copyright infringement when the defender proves he had no knowledge of the previous work (likely because it was an unknown PDF on DMsguild he never bought) and can claim the reproduction was fortuitous. Given the really low odds of it happening, and the good possibility of the defendant proving he had no knowledge of the former work (and the fact that he didn't have to have seen the first work to produce the second], the defendant would be right. In other jurisdictions, both plaintiff and defendant would sort it out through duel to the death. [Standard disclaimer for legal advice on the Internet: there is no one and only solution depending on where you are].
In a totally unrelated topic, whenever I read about flan god, what I imagine is this:
Last edited:
AnotherGuy
Hero
I haven't had time to read all the posts, just a noob query - does the software give one limited opportunities per day unless one subscribes? I have never used DALL-E 3 and I've only fiddled with ChatGPT once. I know! 
Anachronism is my thing.
Anachronism is my thing.
So far, it's mostly available through Bing, so you get a 100 "boosts" that can generate an image in a reasonable time. When you are out of boost, the load on the system is currently such that you need to wait hours for an image. Or overnight, even. Boosts regenerated weekly, then daily, then they got overwhelmed with demand and couldn't cope so they started to hand out boost by pack of 25 when the wait time got too much... Right now it's a mess but one can definitely generate a handful image as you can see in this thread.
I have never used DALL-E 3 and I've only fiddled with ChatGPT once. I know!
Anachronism is my thing.
Dall-E 3 through Bing is limited: you can only generate square images. Since you mention Chat-GPT, there is another way, that is currently being rolled out, through a paid suscription to openAI chat-gpt plus. You get more freedom in your choice of size, you can (supposedly) interact with chat-gpt to make him change an image, but on the other hand it seems to be even more child-friendly as the Bing interface [ie, "beware of those ankles of doom" or "no I won't display Ramses's sarcophagus because he's a political leader and I don't allow political cartoons"]
But the speed of the roll-out is unknown. And it's 25 tokens per 3 hours, so you're still limited even if you pay (20 USD/month).
pukunui
Legend
Been messing around with making character art for my all-gnome Eberron party.
First up is Catnip Evergreen. She's a bow-wielding ranger, as you might expect. However, I'm not even going to go there with the whole bow thing, so I thought I'd try to depict her in her brand-new glamerweave gown that's meant to look like Katniss' "girl on fire" dress. She's also a swarmkeeper ranger, so there are lots of butterflies. With input from my daughter, who is playing her, this is the best we could do with the boosts I had remaining:
Ideally more of the dress would have been on fire, and there would have been more butterflies, and you would have been able to see her feet, but my daughter decided that this would suffice.
Next up is Nissa Nackle. She is a vengeance paladin and templar of the Silver Flame. I had a number of different iterations that were quite cool, but her player settled on this one:
Some of the other contenders:
I've also been working on Dimble Ironhide, an older gnome artificer. His player went with the Ironman-esque armorer subclass, so I put that in one prompt and not in the other. This is what we got:
Next is my wife's PC, Glissando Lyrriman d'Sivis. He's an eloquence bard, and he has a magic lute. My wife liked the face on this one the best, although obviously that's a nutty musical instrument. It also seems impossible to get the dragonmark tattoo to just be on one hand.
The mouse is just there in an attempt to get the full body view.
Last but not least is my youngest daughter's character, Sunny Glitterfall. She is a mix between an elf from a D&D graphic novel and Holga from the D&D movie. She was permanently enlarged to human size by the Mourning, which also led to her becoming a wild magic barbarian. (She's also a bit like Unikitty - she is all sunshine and rainbows on the surface and unbridled rage underneath.) My daughter has gone to sleep, so I'll have to show this to her in the morning, but I think this one is pretty accurate:
First up is Catnip Evergreen. She's a bow-wielding ranger, as you might expect. However, I'm not even going to go there with the whole bow thing, so I thought I'd try to depict her in her brand-new glamerweave gown that's meant to look like Katniss' "girl on fire" dress. She's also a swarmkeeper ranger, so there are lots of butterflies. With input from my daughter, who is playing her, this is the best we could do with the boosts I had remaining:
Ideally more of the dress would have been on fire, and there would have been more butterflies, and you would have been able to see her feet, but my daughter decided that this would suffice.
Next up is Nissa Nackle. She is a vengeance paladin and templar of the Silver Flame. I had a number of different iterations that were quite cool, but her player settled on this one:
Some of the other contenders:
I've also been working on Dimble Ironhide, an older gnome artificer. His player went with the Ironman-esque armorer subclass, so I put that in one prompt and not in the other. This is what we got:
Next is my wife's PC, Glissando Lyrriman d'Sivis. He's an eloquence bard, and he has a magic lute. My wife liked the face on this one the best, although obviously that's a nutty musical instrument. It also seems impossible to get the dragonmark tattoo to just be on one hand.
The mouse is just there in an attempt to get the full body view.
Last but not least is my youngest daughter's character, Sunny Glitterfall. She is a mix between an elf from a D&D graphic novel and Holga from the D&D movie. She was permanently enlarged to human size by the Mourning, which also led to her becoming a wild magic barbarian. (She's also a bit like Unikitty - she is all sunshine and rainbows on the surface and unbridled rage underneath.) My daughter has gone to sleep, so I'll have to show this to her in the morning, but I think this one is pretty accurate:
Last edited:
Composer99
Hero
Apropos of the limits and issues with the technology, this just came across my Facebook feed this morning: Microsoft Bing AI Generates Images Of Kirby Doing 9/11
