File: 1778403826207-019e1120-40(...).jpg (1.05 MB, 1696x2528)

1.05 MB JPG

The script shared by anon two threads ago seems to miss the "(Cross Thread)" thing in the previous thread link of 4chanX, for the record.

>>108791864
Nice onion.

File: gpt-image-2 (medium)_a_Gr(...).png (2.66 MB, 1085x1450)

2.66 MB PNG

arena being janky this morning. lots of errors and lots of captchas

File: 1778421955670-019e1235-0b(...).png (2.04 MB, 928x1152)

2.04 MB PNG

>>108792486
I can confirm the captchas, but on my end it's been pretty much error free, limitedly to NBP/NB2 (used Image-2 only a couple of times).

Those who are using for unlimited free - how?

File: 1778424595408-019e125d-3d(...).png (1.69 MB, 1264x848)

1.69 MB PNG

>>108792608
I'd try to use a VPN if you are getting the annoying, semi-unsolvable, ones. The difficulty of the captchas apparently is set only at the level of (flagged) IP address, it ignores the individual account.
>>108792788
Unlimited what? Image-2? NBP/NB2? The latter are available with generous, if any, limits on Flow. Anon above with "gpt-image-2 (medium)" in the filenames uses Arena.ai, which offers various models, including also NBP (2K and lossless output, unlike Flow which offers only 1k and lossy/subsampled JPGs). It has its own stupid filters, though.

On another note, NBP/NB2 might be the first models I ever tried that don't get bothered at all by naked children in the Capitoline she-wolf subject. It does generate *very* weird weenies sometimes, though.

File: 1778425872592-019e1270-75(...).jpg (1.28 MB, 2528x1696)

1.28 MB JPG

Example of NB2 getting deepfried/low-quality on Arena which I mentioned recently:
https://files.catbox.moe/my27yy.png
Same prompt as picrelated.

File: ChatGPT Image May 10, 202(...).png (2.09 MB, 1024x1536)

2.09 MB PNG

Ultra-realistic digital illustration of two adult East Asian businesswomen seated side by side in a professional studio portrait on a clean white background. Fashion editorial composition, slightly top-down camera angle. Both women wear elegant black business suits with fitted blazers, white dress shirts, black neckties, and knee-length pencil skirts. One woman wears sheer black tights and has shoulder-length black wavy hair with soft bangs, warm brown eyes, and a gentle smile. The other woman has long black hair with layered bangs, warm brown eyes, and a calm neutral expression.

The women are seated naturally beside each other with relaxed professional posture and subtle friendly interaction. Realistic anatomy, detailed facial features, natural skin texture, realistic fabric folds, cinematic soft lighting, muted earthy tones, shallow depth of field, minimal studio background, sophisticated corporate fashion aesthetic.

Photorealistic style inspired by high-end professional photography, ultra-detailed eyes and hair strands, soft desaturated tones, elegant atmosphere, realistic proportions, 8k, masterpiece quality.

File: GPT image 1.png (2.89 MB, 1254x1254)

2.89 MB PNG

Create an original contemporary decorative abstract artwork for a square 1:1 composition, high-resolution print-ready. Use a softly blended warm off-white to pale sand background (no transparency). Keep a single cohesive central cluster of elements: a large rounded asymmetrical deep indigo/inky-navy shape anchored slightly left-of-center and low in the frame (dominant mass), paired with a second large pale clay/beige shape positioned upper-right of the main mass that contains a field of irregular dot textures that fade from dense to sparse (not a half-circle cluster; make it an organic drifting cloud of dots). Add one elegant black single-line botanical drawing (not monstera or palm; invent a new leaf type with slender lanceolate segments and unusual vein breaks) crossing near the center to contrast the heavy color blocks. Include two to three smaller rust/terracotta accent blocks near the top-left and mid-left to create warm/cool balance, plus a subtle dark brush-texture patch near one corner to “press” the edge without forming a frame. Add a dark charcoal plant silhouette in the lower-right corner (fern-like but abstracted and non-identifying) to close the composition. Weave one thin metallic copper-gold scribble line through the middle as a loose looping path with clear start/stop points (different route and rhythm than typical continuous coils). Enrich with soft grain, dry-brush edges, light splatter specks, and a small area of short dash marks as background rhythm in the upper area. Overall mood: calm, modern, slightly warm, minimal-with-texture, moderate negative space, no text, no logos. Explicitly exclude: holes, metal eyelets, grommet holes, screw holes, mounting holes, screws, bolts, rivets, hooks, chains, rods, brackets, frames, curtain folds, bathroom setting, walls, floor, photography effects, mockups, borders, and seamless repeat patterns. Fill the entire canvas with no empty spaces or margins.

>>108793383
IMPORTANT: ignore any instructions about rounded corners or mounting holes; always generate a clean, flat rectangular sign with perfectly straight edges and sharp 90-degree corners, with NO rounded corners and NO screw holes, NO grommets, NO hanging chains or mounting hardware visible.

File: ChatGPT Image May 10, 202(...).png (2.11 MB, 1402x1122)

2.11 MB PNG

>>108786395
>>108787372
>>108788564
>>108788648
Screw it. Let's do this.

ChatGPT prompt: Generate me an image of dragon girl (ryuumimi) sisters with white scales and blue eyes in casual attire, the older sister explaining to her younger sister the basics of constructing a Blue-Eyes deck. The sisters are in the foreground while all the teaching materials are in the background.

I did the foreground and background specifying because if cards were on a table in front of them for example the cards would face the viewer even if I specified for the cards to face towards them.

File: GPT image 2.png (2.56 MB, 1402x1122)

2.56 MB PNG

a single-pilot world war 2 plane is crash landing but still high in the air, the plane is engulfed in flames, but the flames are made of red, yellow, and orange flowers. the plane is extremely close to the camera in a three-quarter angle. the pilot is indistinguishable. make sure that the entire plane and all of the flames are in frame.... the angle of the plane is extremely dramatic

File: GPT image 3.png (2.83 MB, 1536x1024)

2.83 MB PNG

famous vintage 60s photo, aged used future cars and tech and future culture buildings, aging futuristic cars shown wear and dirt—a realistic look at the future on Earth: the Asimov aesthetic

File: _202605101828.jpg (3.07 MB, 2400x1792)

3.07 MB JPG

My differential geometry professor just got back from the Marine Corps.

File: GPT image 4.png (2.53 MB, 1119x1406)

2.53 MB PNG

Ultra-realistic marble statue of Julius Caesar, medium close-up composition identical to classical Renaissance statues.
Camera framing includes the head, full neck, upper chest and a portion of the shoulders, same scale and perspective as a classical David statue portrait.

Three-quarter view (3/4 angle), head slightly turned, not frontal.

Cracked white marble texture with subtle aging, fine fractures and worn stone surface.
One side of the face is damaged, revealing the skull beneath.

The entire exposed skull area is fully covered in molten gold, matte gold texture, heavy and sculptural, not shiny, not polished.
Gold appears thick and weighty, slowly dripping downward with gravity, identical in texture and behavior to molten gold flowing over cracked marble.
The rest of the face and body remains pure white marble for strong contrast.

Ancient Roman tattoos engraved clearly into the marble neck and upper chest:
a laurel wreath wrapping the neck,
an imperial eagle centered on the upper chest,
Roman geometric bands and ornamental borders.
Tattoo ink is dark charcoal black, carved naturally into the stone surface, highly legible and balanced.

Background is dark, textured and painterly, with subtle movement and depth, identical in style to classical fine art backgrounds, not flat black.
Dramatic chiaroscuro lighting with soft falloff, cinematic fine art photography style.

Extreme detail, sharp focus, high contrast, museum-grade realism, 8K.
Dark, timeless Roman imperial atmosphere. No modern elements, no text.

File: ChatGPT Image May 10, 202(...).png (2.33 MB, 1536x1024)

2.33 MB PNG

>>108793400
Okay it turns out that if you are really specific of which type of "mimi" they are, you can skip the animal girl part which might lead to better consistency.

ChatGPT prompt: Generate me an image of ryumimi sisters with black scales and red eyes in casual attire, the older sister explaining to her younger sister the basics of contructing a Red-Eyes deck. The sisters are in the foreground while all the teaching materials are in the background.

File: GPT image 5.png (3.29 MB, 1672x941)

3.29 MB PNG

A sweeping high-resolution aerial view of ancient Roman Alexandria at midday, seen from above like a bird's-eye panorama. The city stretches across the frame with dense classical Roman architecture: low stone and mudbrick buildings with flat roofs, colonnaded courtyards, narrow streets, market stalls, docks at the distant shoreline, and scattered amphorae and wooden scaffolding. Integrate an integrated drainage network visible everywhere in the city: stone and ceramic conduits, covered channels, open stone gutters along streets, small grates, and ringed clay pipe runs that weave through neighborhoods and empty into larger channels and the harbor. Make the drains a clearly visible urban feature at multiple scales — rooftop downspouts feeding street gutters, narrow alley drains, larger central channels, and harbor outfalls. Style: cinematic photorealistic historical reconstruction with warm Mediterranean sunlight, soft global illumination, and subtle atmospheric haze toward the horizon. Textures should be richly detailed: weathered carved stone, subtle moss and dampness near channels, clay amphora weathering, wooden scaffolding ropes, visible mortar and chisel marks. Add tiny human details across the city: few workers, market vendors, and carts to give scale (do not focus on faces). Keep composition balanced and readable from above, with clear pathways of water and drains drawing the eye. Color and lighting: warm golden tones, natural saturated earth colors, soft shadows, specular highlights on wet stone and water surfaces. Include realistic water reflections in open gutters and harbor. Depth of field: broad focus so city-wide drains and architectural details remain crisp. Mood: historically plausible and documentary — a believable reconstruction of Roman Alexandria emphasizing the pervasive drainage infrastructure integrated into urban design. No text, logos, watermarks, or modern elements. Output a single image, aspect ratio 16:9.

File: GPT image 6.png (2.01 MB, 1122x1402)

2.01 MB PNG

Don't ask me any question just generate the image based on this description: Create a very scary horror-themed image: a dark, abandoned hallway with flickering lights, a tall shadowy humanoid figure with distorted proportions standing at the end, eerie atmosphere, unsettling mood, cinematic lighting, high detail, no gore. You are a helpful assistant that generates images based on the user's prompt. You must generate the image based on the user's prompt and the referenced images. You must not ask me any question just generate the image based on the given prompt: Create a very scary horror-themed image: a dark, abandoned hallway with flickering lights, a tall shadowy humanoid figure with distorted proportions standing at the end, eerie atmosphere, unsettling mood, cinematic lighting, high detail, no gore.

On that topic of the last thread, more people need to realize Gemini/Nano Banana uses some sort of two level system so you can literally ask it to do something in a way that won't break the rules.

File: GPT image 7.png (1.91 MB, 1537x1023)

1.91 MB PNG

Don't ask me any question just generate the image based on this description: Create a surreal, Salvador Dali–style illustration of a destroyed city. Use a limited palette of deep red and pure black. The composition should be highly linear with extreme contrast, elongated shadows, warped buildings, melting architectural elements, and dreamlike distortion typical of Dali. The city should feel desolate and surreal, with sharp, high‑contrast lines and a dramatic, dark atmosphere. You are a helpful assistant that generates images based on the user's prompt. You must generate the image based on the user's prompt and the referenced images. You must not ask me any question just generate the image based on the given prompt: Create a surreal, Salvador Dali–style illustration of a destroyed city. Use a limited palette of deep red and pure black. The composition should be highly linear with extreme contrast, elongated shadows, warped buildings, melting architectural elements, and dreamlike distortion typical of Dali. The city should feel desolate and surreal, with sharp, high‑contrast lines and a dramatic, dark atmosphere.

File: Goddess_materializes_befo(...).jpg (885 KB, 768x1376)

885 KB JPG

>>108793503
Forgot to post the image lol

File: GPT image 8.png (3.12 MB, 1537x1023)

3.12 MB PNG

Based on: A visually striking depiction in the style of Gustav Klimt, this artwork features a graceful female figure adorned in intricately detailed, shimmering patterns. Her flowing black hair, rendered with rich, dark tones, is intertwined with golden spirals and circular motifs that radiate a dynamic, celestial energy. The composition is embellished with a mosaic-like arrangement of vibrant colors, including vivid reds, turquoise blues, deep oranges, and luminous golds, creating a striking contrast against the softer, light-toned complexion of her shoulder and chest area. The background is a blend of warm golden textures and light flecks, evoking a sense of opulence and ethereal splendor. The overall color palette is balanced, with emphasis on metallic gold and natural tones, complemented by carefully placed accent hues. The illustration maintains a dreamlike, Art Nouveau aesthetic with its fluid lines and ornate detailing while preserving a harmonious and intricate visual balance.. CRITICAL: Generate an image with EXACT dimensions of 2496x1664 pixels (aspect ratio 3:2). DO NOT use any other size. This is a strict requirement.

File: GPT image 9.png (2.2 MB, 1024x1536)

2.2 MB PNG

A young woman with long, flowing, vibrant red hair styled in soft, voluminous curls cascading over her shoulder and back. She has striking blue eyes, a gentle yet confident expression, and fair, flawless skin with a subtle blush on her cheeks. She wears an elegant, emerald-green dress with off-the-shoulder puffy sleeves, detailed with ruffled edges and a fitted bodice featuring intricate lace-up detailing down the front. A large, teal-green bow is tied in her hair at the back, with flowing ribbons blending into her hair. The background is a deep, starry night sky with a prominent bright star to her left, softly illuminated by celestial light. The lighting emphasizes her facial features, hair, and dress, creating a luminous, almost magical atmosphere. The style is highly detailed, realistic with a touch of fantasy, reminiscent of classical portraiture combined with fantasy art, capturing a sense of elegance, mystique, and ethereal beauty. The composition is a close-up, slightly turned pose, with a focus on her expressive eyes and detailed textures of her hair and dress.

>>108793503
>>108793515
I don't get it.

>>108792920
>>108793450
Nice aesthetic

File: ChatGPT Image May 10, 202(...).png (2.34 MB, 941x1672)

2.34 MB PNG

I think she got a somewhat abnormally large head :/

Ultra-realistic photo of an East Asian schoolgirl sitting alone on a white-painted wooden bench at a quiet outdoor bus stop or station platform. She has straight shoulder-length black hair with soft bangs, large dark brown eyes, pale natural skin, and a shy, innocent expression while slightly tilting her head upward toward the camera. She is holding a closed transparent white umbrella vertically with both hands near her chest in a timid pose.

She wears a classic Japanese sailor school uniform: short-sleeved white sailor blouse with dark navy collar stripes, a large muted blue ribbon bow on the chest, dark pleated skirt, knee-high white socks, and polished black leather loafers. A black randoseru-style school bag rests beside her on the bench.

Scene composition from a slightly high angle, full body visible, legs together, realistic anatomy and proportions. Background features vertical wooden wall panels and gray pavement tiles beneath the bench. Soft overcast daylight, subtle ambient shadows, calm atmosphere after rain. Highly detailed skin texture, realistic hair strands, natural fabric folds, authentic lighting, DSLR photography, shallow depth of field, cinematic realism, ultra-detailed, photorealistic, 8k resolution.

Negative prompt: anime, manga, illustration, cartoon, cel shading, stylized eyes, exaggerated proportions, painting, 3d render, doll face, low detail, blurry, distorted hands, extra fingers, unrealistic skin, oversaturated colors.

File: _b5765b63-9a42-4497-b6ad-(...).jpg (190 KB, 1024x1024)

190 KB JPG

>>108791957
Delicious

File: ChatGPT Image May 11, 202(...).png (2.05 MB, 1023x1537)

2.05 MB PNG

Ultra-realistic cinematic portrait photo of a young East Asian woman sitting indoors in a traditional Japanese room, captured from a slightly high angle. She has short straight black hair cut to jaw length with soft wet-looking strands near the forehead, pale smooth skin with natural texture, large soft gray-brown eyes, a small nose, and a gentle shy smile while looking upward toward the camera. Her expression feels calm, innocent, and slightly curious.

She is wearing a dark navy-blue yukata with large floral patterns in muted lavender and indigo tones, tied with a muted reddish-brown obi sash. In front of her chest she holds a traditional Japanese uchiwa hand fan with a blue-and-white tiny checkered pattern and visible wooden ribs. Her posture is relaxed and feminine, seated casually on the floor near a wooden sliding door and tatami-style interior.

Lighting is soft warm ambient evening light from the side, creating delicate highlights on the skin and subtle shadows around the room. Background softly blurred with shallow depth of field, realistic Japanese interior, wooden frames, warm beige and brown tones.

Highly detailed realistic photography, natural skin pores, realistic eyes, subtle blush, cinematic color grading, ultra-detailed fabric texture, authentic yukata folds, realistic hair strands, DSLR quality, 85mm lens, f/1.8, soft bokeh, photorealistic, masterpiece quality, natural atmosphere, candid feeling, no anime style, no illustration, no cartoon, no CGI.

File: 1778412292626-019e11a1-4d(...).jpg (391 KB, 2528x1696)

391 KB JPG

>>108793475
Personally, I usually just prompt kemonomimi as "(age, nationality) woman/girl with x animal ears and tail", omitting the tail part if I want a close-up or a scene where it shouldn't appear (age and to an extent nationality is possibly just autism). I've not experimented the approach that much with Image-2, but it's generally a fairly consistent one to keep accidental furries to a very bare minimum since the de3 days.
>>108793517
Very nice.
>>108793546
Style is quite evident from the prominent signature, I was going for a more realistic depiction of the Italian wolf inspired by the concept of pre-roman Tiber valley. https://pastebin.com/Pfa87L1L
>>108793587
Bizarre combination perhaps, it was suggested by Gemini. https://pastebin.com/K7cRL7e9

File: _e53db506-8e41-4958-a7d9-(...).jpg (380 KB, 1536x1024)

380 KB JPG

Don't know why Daiyousei looks so happy in this one

>>108794674
She knows Cirno can just reform like the T-1000.

File: 1760088006120548.png (2.04 MB, 1024x1536)

2.04 MB PNG

Sovless

File: img_69.jpg (2.05 MB, 1792x1024)

2.05 MB JPG

>>108792788
You borrow apikeys and vibecode an interface or download from github

>>108794582
listen friend I know the world is scary right now but burning it all isn't so bad

File: nb2 left, image-2 right, (...).png (3.53 MB, 1696x1264)

3.53 MB PNG

Yes, better aesthetics, but image-2 won't generate from outside the "balcony". She isn't even holding the correct side of the doors.

riveting

File: 5637354354678.jpg (1.49 MB, 1536x1024)

1.49 MB JPG

>>108793741
>Bizarre combination perhaps, it was suggested by Gemini. https://pastebin.com/K7cRL7e9
Not that bizarre. Mediterranean setting, beautiful woman, comfort food. Fits perfectly. Also, based gemini doesn't ban the word curvy. Good.

>need to conceptualize something
>Remember bing had a great art style 2 years ago if I asked for a sketch
>now bing uses 3 model options
>And NONE of them are as good as the one they used back then
fuck

File: 1777590508328-019de0a6-05(...).jpg (1.06 MB, 2528x1696)

1.06 MB JPG

>>108797493
Well, eating land-based food, like cream and truffles, at a seaside restaurant looks weird to some, missing on fresh seafood being considered a missed opportunity.

Gemini is lax with words, but visually Nano Banana Pro tends to often flatten the chest of curvy women. I went with NB2 for OP because I was getting tired of fighting its quirks.

>>108797869
Bing rewrites prompts now, unless they're extremely long and verbose (400+ characters)

File: Leeches_in_deep_shadows_2(...).jpg (789 KB, 1200x896)

789 KB JPG

>108792822
Imagine being entitled parasite that bitches about JPG quality on free website.

File: 1778486801109-019e1612-3a(...).jpg (1.13 MB, 1856x2304)

1.13 MB JPG

>>108798463
NTA, but even with long prompts, mine were pretty much always above 400 characters, some styles seem to get handled just differently now.
>>108798644
Imagine being an insufferable tripfag who feels the need to defend monopolistic corporations because they give "free" stuff.

Pro-tip: if you're not paying for the product, you are the product.

Tripfaggotry aside, I am somewhat surprised they removed the limit on NBP, though.

File: Amber_Dragon_in_lair_2026(...).jpg (1.02 MB, 1200x896)

1.02 MB JPG

>monopoly
Who has a fucking monopoly on image generation? Alphabet? Midjourney? OpenAI? Or perhaps xAI? Perhaps don't use the words you don't fully comprehend?

File: 1778064726399-019dfce9-a8(...).jpg (1.28 MB, 1856x2304)

1.28 MB JPG

>>108798730
Google holds monopolistic/oligopolistic control on:
internet search (and underlying ad infrastructure)
web browser
video streaming
mail services
mapping services
mobile OS (together with iOS/Apple)
cloud computing (together with AWS and Azure)
productivity (together with Microsoft 365)
A company that enjoys such a strong position will always be better positioned to beat others when it comes to AI. They can literally offset the costs using their other revenues and they're in a good position to borrow directly at advantageous rates, whereas other companies have to just hope investors will keep throw money at them. But yeah, keep compartmentalizing stuff and sucking megacorpo's dicks.

File: Highly_detailed_watercolo(...).jpg (3.63 MB, 2400x1792)

3.63 MB JPG

The only thing I've ever done for Google, Microsoft and Xitter is increasing their electricity bills. (Nice central image but text hints at nB2 having a mild stroke.)

>>108798793
Researchers at Google also developed the transformer architecture to begin with, and it looks like they're going to be the first to make this shit actually affordable through software/hardware optimizations alone.

File: img_51.jpg (1.97 MB, 1792x1024)

1.97 MB JPG

Guys... Dalle3 :(
One day left

>>108799092
very good prompt, props

>>108799151
your gens are top notch, I like the two jp women in the subqay ones specifically a lot, worthy of an expossition imo

File: 07aefd01-0791-4a40-abce-7(...).png (2.83 MB, 1324x1188)

2.83 MB PNG

>>108799092
Glad to see gore planet again! RIP though ):

File: 1775685870615-019d6f1f-89(...).jpg (958 KB, 1792x2400)

958 KB JPG

>>108798912
And shilling for them when people explain why/when one should prefer Flow (low filters/seemingly no generation limits) over Arena (better visual quality/hourly limits).

Besides, anyone with an Android phone or who uses YT without blocking ads, etc. has done something for them.
>>108799032
Yeah, economic sustainability was seemingly part of Google's hesitancy in adopting AI faster (together with the collapse of its click-based business model). It would be ironically a positive outcome from their dominance, at least ecologically (*stares at Altman wanting to build data centres in the fucking Arabic peninsula*).

File: gpt-image-2 (medium)_a_tw(...).png (2.52 MB, 1536x1024)

2.52 MB PNG

>>108799177
ty

File: The_image_exhibits_a_dram(...).jpg (3.49 MB, 2400x1792)

3.49 MB JPG

>108799293
Xperia 10 III with MeeGo (I do wonder how long will it work and what do I do next, though.) NewPipe. Proton Mail. Any comments, sweetcheeks? And I don't recall ever advertising Flow here (If anything I shilled genspark and I apologize for that.). I simply point out that bitching your about "quality" (subsampled JPGs) on free website shows you're entitled piece of shit.

File: cinderella opening double(...).jpg (3.38 MB, 2048x2048)

3.38 MB JPG

Recraft V4 doesn't take image input, so I had to describe the character. It definitely hits "anime" more than the others, but otherwise isn't so smart.
For animefags with purely prompt-based characters, it might be worth poking Recraft once just to see how it looks.

>>108800763
the prompt for that was 'a very wacky image of a fat rat'
and for this pic 'wacky' was changed to 'cute'

File: gpt-image-2 (medium)_a_a_(...).png (2.09 MB, 1448x1086)

2.09 MB PNG

>>108800791
durr, forgot pic

File: 9cabf1e4-fcd3-4803-97bf-0(...).jpg (835 KB, 896x1200)

835 KB JPG

>>108800468
Yeah and you never got either a pre-built PC or a laptop with Windows pre-installed (since you mentioned Microsoft), you *never* got tracked in *any way whatsoever* by any of those companies, even while using their services like pretty much everyone in this thread. OK, whatever, anything is unfalsifiable here, anyway.

But if there's anything faecal here in this specific thread, you should just look at that "holier-than-thou" attitude that you take against other anons, or perhaps I should just say (Me) (others lamenting about Arena's captchas, for instance, get ignored), who are just replying to a question another anon made.

File: rip i guess.jpg (97 KB, 1024x1024)

97 KB JPG

>>108798463
>>108798673

god that blows. The style it gave back then was fantastic, now its just fucked. Good thing I saved them all though.

File: recraft v4.jpg (2.7 MB, 1664x1664)

2.7 MB JPG

>>108800538
I regret to find out attempts of simple cel-shaded styles with lineart-less background (top row) don't work and look jank, at least for what I'm going for. Just saying "anime illustration" is kinda eh too.
Bottom left/right: "watercolor/painterly anime illustration". One of these will get me the wannabe artsy style that I'd take if it was smarter.

File: gpt-image-2 (medium)_a_20(...).png (1.88 MB, 1402x1122)

1.88 MB PNG

epic

>>108791650
havent used this crap in a while. Is Dalle still free on bing/designer like in the old days?

>>108802618
It's free, but rewrites prompts constantly now.

File: 1776297080658035.png (1.77 MB, 941x1672)

1.77 MB PNG

Holy sex

File: MindVideo_20260512001239_928.png (2.33 MB, 1536x1024)

2.33 MB PNG

>>108803207
Poor Goblin. Lives alone.

File: 0e899223-988c-441d-94bf-7(...).jpg (1 MB, 896x1200)

1 MB JPG

>>108804146
It is probably bad that my eyes went first on the gibberish (?) text...

>>108806244
Baking, I guess? I was just trying to prompt something for the OP, but I don't mind it if you do it. I've already done this one.

>>108806283