📢 로그인 문제 해결 안내 — 오늘 일부 사용자에게 PixAI 계정이 예기치 않게 로그아웃되는 현상이 발생했습니다. 해당 문제는 서버 측 오류로 확인되었으며, 현재는 정상적으로 해결되었습니다. 로그아웃된 경우 앱 또는 브라우저를 다시 실행한 뒤…
자세히 보기
PixAI에서 프롬프트는 기본적으로 공개되지만, 창작자는 자신이 올린 작품마다 프롬프트를 숨길 수 있습니다. 이 작품의 창작자는 프롬프트를 숨기기로 선택했기 때문에, 이미지를 생성하는 데 사용한 단어는 창작자 본인만 볼 수 있습니다. 프롬프트를 숨겨도 그 외에는 달라지는 점이 없습니다. 작품은 그대로 공개되어 있고, 평소처럼 좋아요, 댓글, 공유를 할 수 있습니다.
작성자에 의해 숨겨짐
모델 및 LoRA 사용
비슷한 작품
자세히 보기
gas-!
gas-!
gas-!
ｍ✚
A deliberately terrible, awkward digital doodle based loosely on the original character and composition.
White background.
Drawn as if a person in the late 1990s or early 2000s used an old PC painting program with only a mouse, no tablet, no pressure sensitivity, no professional drawing skills, and very little control over the cursor.
The result should be strangely recognizable but fundamentally wrong.
The character should vaguely resemble the original subject, but many things should be subtly misplaced, distorted, simplified, or awkwardly connected. The drawing should constantly feel like it is almost correct and then slightly misses the intended shape.
EXTREMELY LOW ARTISTIC QUALITY.
Awkward mouse-drawn lines.
Uneven outlines.
Accidental-looking corners.
Wobbly curves.
Abrupt changes in line direction.
Lines that overshoot and stop too late.
Lines that stop too early.
Slightly misaligned facial features.
Eyes that are almost symmetrical but not quite.
A mouth that is slightly too high or too low.
Hair that vaguely follows the original hairstyle but has clumsy, lumpy shapes.
Hands that are simplified into strange mitten-like forms.
Feet that are recognizable as feet but proportionally wrong.
Arms and legs that connect to the body at slightly incorrect locations.
Joints that bend in ways that are technically possible but visually uncomfortable.
Anatomical proportions that are close enough to identify the character but obviously incorrect.
Do not make the anatomy elegant.
Do not correct the anatomy.
Do not improve the proportions.
Do not make the pose graceful.
The pose should have the peculiar stiffness of someone trying to copy a complicated illustration while lacking the skill to understand its underlying construction.
The character should look as though the artist remembered the original image imperfectly and attempted to reproduce it from memory.
Keep a few recognizable characteristics from the original:
the overall silhouette,
the long hair,
the reclining sideways composition,
the approximate placement of the face,
the approximate placement of the legs and feet,
and the general clothing shape.
But deliberately introduce small errors everywhere.
The face should be especially funny:
approximately the right shape, but slightly crooked.
One eye subtly larger than the other.
Eyes positioned a little too high or too far apart.
Nose represented by an unnecessary tiny line.
Mouth slightly misplaced.
Expression vaguely similar to the original but unintentionally awkward.
Hair should be represented with a small number of crude, chunky shapes rather than carefully drawn strands.
Some hair shapes should terminate strangely.
Some strands should overlap incorrectly.
The overall hairstyle should remain recognizable while looking poorly copied.
Use primitive shapes whenever possible:
rough circles,
lumpy polygons,
uneven rectangles,
simple curved strokes,
badly joined segments.
The drawing should look as if every shape was manually placed one pixel or one mouse movement at a time.
PIXEL QUALITY:
Very low-quality raster appearance.
Small canvas feeling.
Visible individual pixels.
Hard pixel steps.
Jagged diagonal lines.
Uneven pixel clusters.
Crude one-pixel corrections.
Occasional isolated pixels.
Slightly inconsistent pixel density.
Primitive fill-tool coloring.
Flat regions with accidental gaps.
Tiny unfilled white spaces.
Small areas where the color spills one or two pixels outside the intended boundary.
Do NOT make it polished pixel art.
Do NOT make it professionally designed pixel art.
Do NOT make it retro-game artwork.
Do NOT make the pixels aesthetically perfect.
Instead, make the pixels look like the unfortunate consequence of drawing badly with an old mouse.
COLOR:
Use a very limited and slightly unpleasant old-computer palette.
Colors should be approximately related to the original image, but not accurately matched.
Some colors should be inexplicably too bright.
Some should be slightly too dark.
Some neighboring colors should clash.
Skin may be slightly too pink or too yellow.
Hair may be slightly too brown or reddish.
Clothing may be almost correct but subtly wrong.
Flat color fills with occasional accidental overlaps.
No sophisticated gradients.
No realistic shading.
No smooth rendering.
No atmospheric lighting.
Shadows should be crude, misplaced, and geometrically awkward.
A shadow may appear in a location that technically makes little sense.
Highlights may appear where the artist apparently thought they looked nice rather than where light would actually reach.
LINE QUALITY:
Use crude one-pixel or few-pixel strokes.
No clean professional lineart.
The line thickness should vary for completely arbitrary reasons rather than according to anatomy, lighting, or depth.
Some contours should be darker.
Some should almost disappear.
Some should have tiny gaps.
Some should be redrawn directly beside an earlier incorrect line, creating a visibly nervous correction.
Occasionally leave the original wrong stroke visible.
Make it feel like:
draw → notice mistake → draw another line next to it → give up.
The image should contain tiny visual mistakes everywhere.
COMPOSITION:
Keep the character large enough to occupy most of the canvas, but do not achieve professional composition.
The figure should be slightly awkwardly positioned.
Some areas should feel too cramped.
Other areas should have inexplicably empty space.
The complete body should remain broadly recognizable, including both feet, but their placement can be slightly wrong.
Do not crop the important body parts in a sophisticated way.
The character should feel as if someone manually tried to fit the entire complicated figure into a small white canvas and barely managed it.
BACKGROUND:
Pure white background.
Absolutely no room.
No wall.
No floor.
No bed.
No scenery.
No decorative background.
No environmental objects.
Only the badly drawn character on white.
The white space should make the low quality even more obvious.
IMPORTANT OVERALL EFFECT:
The image should create a very specific feeling:
“It is obviously trying to depict the original image, but whoever drew it simply cannot draw.”
Not abstract.
Not surreal.
Not intentionally grotesque.
Not horror.
Not distorted for shock value.
Just painfully mediocre.
A strangely sincere amateur reproduction.
The viewer should be able to say:
“Yeah, that's probably supposed to be the same character.”
while immediately noticing:
“but almost everything is slightly wrong.”
Make the mistakes accumulate.
The face is a little wrong.
The hair is a little wrong.
The torso is a little wrong.
The hands are a little wrong.
The feet are a little wrong.
The perspective is a little wrong.
The colors are a little wrong.
The linework is a little wrong.
The spacing is a little wrong.
But nothing is wrong enough to completely destroy recognition.
The image should feel like a low-skilled person desperately attempting to reproduce a sophisticated illustration with nothing but a mouse and an ancient painting program.
No polished aesthetics.
No professional composition.
No beautiful rendering.
No smooth anatomy.
No sophisticated lighting.
No cinematic quality.
No high-end illustration finish.
Maximum awkwardness.
Maximum amateurishness.
Maximum visual incompetence.
The final result should be charming precisely because it is so poorly made.
It should look like a treasured, embarrassingly bad old computer doodle that someone made with complete sincerity.
Ｍｒ．✚ナタ 'ﾉ
추천 사용자
당신에게 추천
yoimiya!
ᯓᯓ 𝙘𝙤𝙘𝙤𝙛𝙡𝙪𝙛𝙛𝙞
Mine / 私のもの
みーちゃん
✦ ₊ ⊹ 🌑
(mai)
❌⃝✦—𝐇𝐄𝐘! 𝐘𝐨𝐮 𝐰𝐚𝐧𝐧𝐚 𝐜𝐨𝐦𝐞 𝐢𝐧?⬩➤“
𝐋𝐢𝐗𝐗𝐗┆秋峠鬼火~👻
the queen
ᯓᯓ 𝙘𝙤𝙘𝙤𝙛𝙡𝙪𝙛𝙛𝙞
the lancer
ᯓᯓ 𝙘𝙤𝙘𝙤𝙛𝙡𝙪𝙛𝙛𝙞
🎵Surtur
helanx
マシュ
Atlas
Sakura Chiyono O: A Springtime Portrait
AfterThoughts
Love live! School idol maid - Hanayo koizumi
AMnatsu
음안
character: eren_yeager, mikasa_ackerman
character: eren_yeager, mikasa_ackerman
Mikasa Ackerman and Eren Yeager are captured in the middle of a desperate ODM maneuver above a ruined city street, with Mikasa much closer to the camera and Eren cutting across the background at a different angle. Mikasa twists her entire body sideways as she launches past the viewer, one arm extended forward with a blade in hand while the other reaches back toward Eren. Her black hair whips violently across the frame and her red scarf streams far behind her in a long sweeping curve. Eren is several meters behind her, angled sharply in the opposite direction as his ODM cables stretch from his gear toward a damaged rooftop, his green jacket and hair blown backward by the speed of the maneuver.
The perspective is extremely dynamic: Mikasa's outstretched hand and blade are large in the foreground while her body rapidly narrows toward the center of the image. Eren appears smaller but perfectly clear behind her, creating a strong diagonal connection between them. Their ODM cables slice through the composition as thin steel lines, crossing the image at different angles and disappearing beyond the frame. One cable passes extremely close to the camera, creating a dramatic foreground line.
The ruined city beneath them is not shown as a wide landscape. Broken rooftops, shattered windows, hanging masonry, and fragments of brick rush diagonally through the lower portion of the composition, giving the impression of immense speed. Several pieces of debris are suspended around Mikasa and Eren, each frozen at a different distance from the camera. Dust catches the sunlight in thin atmospheric streaks.
Mikasa's Survey Corps uniform is sharply detailed and windswept, with the cropped brown jacket open over the fitted uniform, harness straps pulled taut across her body, white trousers, tall boots, and the Wings of Freedom insignia clearly visible. Eren wears his recognizable Survey Corps uniform and green jacket, with his harness, ODM equipment, and boots rendered with precise mechanical detail.
A subtle graphic-design layer is integrated into the composition without overpowering the action: Wings of Freedom emblem are overlaid across the sky, Thin white directional lines follow the trajectories of Mikasa and Eren, while small restrained military-style markings and typography sit along the edges of the frame. The graphic elements reinforce the movement rather than becoming random decoration.
The entire image is frozen at one extraordinary instant: Mikasa's scarf curves through the air, Eren's jacket flares behind him, cables are pulled completely taut, debris hangs in the air, dust streams diagonally through the sunlight, and both characters are caught at different points of the same sweeping maneuver. Nothing feels static.
Painterly digital anime illustration, expressive textured brushwork, cinematic motion frozen in a single frame, extreme wide-angle perspective, dramatic foreshortening, strong diagonals, layered depth, detailed characters, flowing hair and fabric, taut cables, airborne debris, warm sunlight cutting through cool gray ruins, muted military colors with Mikasa's red scarf as the strongest accent, sophisticated composition, subtle graphic design, powerful silhouettes, intense expressions, visceral sense of speed, beautiful atmospheric lighting, energetic, cinematic
음뭐지 안되서 다시 부계만듬
🍧💙
Janet reno
໒꒱· ﾟ◼️
生魚2(Raw Fish 2)
Gojo & Geto's Playful Selfie! ✨
Momo
The Empress of the Violet Abyss 👑💜
wisely
👻 💬 📂
(mai)
⋆⁺₊⋆🪞🕯️⋆⁺₊⋆
(mai)
(-_-)🚬⚡︎
(mai)
(っ.❛ ᴗ ❛.)っ🫧🐟
(mai)
🌩️🐾(⸝⸝ᵕᴗᵕ⸝⸝)⛳️
(mai)
❓
(mai)
耍车
星空sky
✧ 🍁 🧣 🍂 ✧
(mai)
(*˘︶˘*).｡*♡ 🍁 ☕
(mai)
댓글