Hooded crowned masked eldritch king stands — AI art by @legion
@legion · 9/29/2026
Hastur.

Tags: eldritch king, dark fantasy art, surreal horror, desolate desert landscape, lovecraftian art, masked monarch, apocalyptic world, twin suns, gothic fantasy, dark surrealism, barren wasteland, cursed ruler, horror illustration, sepia fantasy art, dead forest scenery

Tags: King in Yellow, Hastur, eldritch horror, Lovecraftian, dark fantasy art, ancient ruins, apocalyptic landscape, cosmic horror, hooded entity, dead city, golden dusk, surreal illustration, crumbling stone architecture, eerie atmosphere, gothic fantasy

Tags: gothic horror, dark fantasy, lovecraftian creature, abandoned cathedral, supernatural entity, spooky atmosphere, stained glass windows, foggy interior, tentacled monster, gothic architecture, haunted church, horror illustration, dark digital art, creepy fantasy, moody horror

Tags: dark fantasy art, surreal horror, swamp entity, moonlit swamp, haunting illustration, eerie creature, gothic fantasy, cosmic horror, spooky landscape, mysterious entity, atmospheric horror, wetland night scene, concept art, dark surrealism, horror artwork
Comments
legion
It’s always done that for me—generating four images from a single prompt without me having to ask for anything extra. If you upload an image to modify, then it gives them to you one by one, but it will do it if you ask; I once uploaded a photo of myself and told it to place me in different settings, and it did so without a problem. Maybe it happened because you used your own characters in the ones you created; for these images, for example, I just used "Hastur" as the prompt and got these four results.
legion
Sure, I told her to make them however she wanted but in different styles—maybe that’s why they turned out better—but maybe if you add too much stuff, it doesn't work as well. Do you use version 4.5 or 5? I think the former leans more towards illustration, while 5 leans more towards realism.
panos
I think it asked me whether to make just one image. I'm not sure. All the ones I did yesterday were with imported characters so it can be as you say. Right now i'm working on the first page of a comic book story, see how that will go.
legion
Great, I hope it goes well. I did a few random pages in different styles and it seems to have gone well—there were a few typos in the text, but nothing major.
charlypalermo
Hahaha, the same thing has been happening to me for a while now... and I keep trying because @legion gets some great images... but I can only get them with GPT
charlypalermo
Espectaculares! Creo la 4 es la que mas me gusta, luego la 1, la 3 y la 2. Pero realmente son grandiosas todas. DOLA es tu amiga, conmigo se porta mal jaja
panos
well, I don't know, my friend. I havn'r checked what version it is. I'll do it today. I will continue using it for images, it is great in that!
legion
Yes, thanks friend. That’s something I like about Dola: sometimes it gives you four images of the same subject, but they’re very different.
panos
Unfortunately, it didn't go well. I gave it averything, imported characters, description of panels, text, and it made huge mistakes, 3 times. 😞
panos
You're wellcome, my friend. Oh, you can ask it to do a prompt in batches of four?