Jeewoo258

Nothing so say - Nothing to show

[ΛRMIN @Arminn_Ai]

3D miniature isometric room with Nano Banana Pro.

Prompt👇

[梦老湿 Dream AI @jackzhang123vip]

真实感哪里来的？真人感哪里来的？不是越干净越真实，有适当的光源，脸部明暗关系的对比，加上一些光线质感，丰富下细节，真人感一下子就会上来
今天分享下光影的使用，可以点赞收藏

[Louise @LouiseVisual]

Inkfleck

--sref 535019022

[simeon-sanai @Naiknelofar788]

GPT image 2

Prompt:
Create a premium vertical 4:5 travel poster for [COUNTRY], illustrated as one elegant continuous hand-drawn line-art composition.

Instead of placing landmarks separately, make them flow organically into one another, as if the entire country was drawn in a single uninterrupted travel sketch. Combine the country’s most recognizable architecture, landscapes, culture, food, transportation, nature, and tiny everyday details into one beautifully balanced scene.

Use bold black ink outlines, delicate cross-hatching, playful imperfect hand-drawn details, subtle vintage-poster typography, and a mostly warm off-white paper background. Add only 2–3 signature accent colors inspired by the country.

Include:

One dominant iconic landmark
A second architectural landmark subtly integrated into the skyline
A famous natural landscape in the background
Local transportation
Traditional clothing or cultural detail
A recognizable food or drink
Small street-life moments
Birds, clouds, plants and tiny decorative elements
Handwritten labels pointing to selected elements

At the top, use a large playful hand-lettered title:

“[COUNTRY]”

Below it, add a tiny elegant subtitle:

“A little world to discover”

Make the composition feel collected, whimsical, editorial, nostalgic and handcrafted, like a beautifully illustrated travel journal rather than a generic tourist poster. Avoid photorealism, gradients, 3D rendering and excessive colors. Clean white space around the illustration, highly detailed ink work, charming imperfections, premium stationery aesthetic, screen-print texture, sophisticated yet playful.

[rui｜∫varts (シュバルツ) @varts_works]

TapNowキャンバス内のMiniMax H3で制作
キャラシート1枚で作った15秒PV👠

キャラシートだけでここまで出来るなら素材作らなくて済む😲みんなもやってみてね〜✨

制作: @TapNow_JP @TapNow_AI #TapNow
衣装: LUCY @ParallaxTalent #ParallaxAgency

プロンプト長いので2つに分けてます🔻

[rui｜∫varts (シュバルツ) @varts_works]

Prompt 2/2
​─────​─────

8.0–10.0s —

Top-down camera begins descending toward her while slowly rotating.

NERA remains almost completely still.

Only her head follows the camera.

She raises one red-gloved hand upward toward the descending lens.

Before touching it, she suddenly lowers her hand.

The camera drops rapidly past her shoulder.

Match cut.

10.0–12.0s —

NERA now walks horizontally across the frame from right to left.

The camera tracks perfectly parallel with her.

Behind her, enormous black letters appear one at a time on the yellow wall:

N

E

R

A

She physically walks past each letter.

Her body temporarily blocks portions of the typography, integrating her silhouette into the graphic design.

On the final “A,” she stops.

12.0–13.5s —

NERA turns directly toward the camera.

The camera suddenly performs a rapid push-in from full body to upper-body framing.

Instead of posing, she takes one deliberate step toward the lens.

A red-gloved finger briefly passes vertically across the center of the frame.

The image visually splits into two offset versions of NERA for several frames.

They immediately snap back together.

13.5–15.0s —

Silence in movement.

Clean symmetrical medium portrait.

NERA stands completely still against pure yellow.

Her black veil moves slightly from a subtle studio breeze.

She slowly tilts her chin downward while maintaining eye contact.

A narrow horizontal shadow passes across her eyes.

As the shadow disappears:

NERA

appears extremely large behind her.

She gives the slightest confident expression.

Instant hard cut to yellow.

END.

VISUAL LANGUAGE:

photorealistic live-action fashion film,
luxury Japanese fashion editorial,
minimal yellow studio,
vivid red × yellow × black graphic palette,
strong negative space,
geometric composition,
controlled studio shadows,
glossy patent leather reflections,
realistic skin,
premium commercial cinematography,
bold framing,
restrained model performance,
sharp rhythmic editing.

MOVEMENT LANGUAGE:

slow confident walking,
sudden precise stops,
controlled head turns,
subtle eye contact,
small deliberate hand gestures,
minimal elegant posing.

NERA should never perform exaggerated dancing or generic influencer poses.

CAMERA LANGUAGE:

extreme wide static composition,
fast lateral tracking,
ground-level tracking,
top-down descending camera,
parallel side tracking,
rapid linear push-in,
precise symmetrical framing.

Transitions should primarily come from:
foreground occlusion,
NERA crossing the lens,
moving shadows,
hard cuts,
camera flashes,
match cuts,
physical camera movement.

Do not rely on generic digital transition effects.

IMPORTANT:

The reference image is a visual identity sheet only.

Do not recreate its multi-view character-sheet layout inside the video.

Do not copy choreography, shot order, transitions, timing structure or compositions from another fashion animation.

The resulting film must feel like an original NERA fashion campaign designed specifically around her red patent outfit and yellow visual identity.

AVOID:

anime,
illustration,
3D character rendering,
character-sheet presentation,
comic panels,
collectible-card montage,
ribbon transitions,
hair-curtain transitions,
eye-detail montage,
generic fashion montage,
cyberpunk,
neon environments,
complex scenery,
heavy particles,
outfit changes,
identity drift,
facial changes,
hairstyle changes,
missing accessories,
unrealistic anatomy,
extra fingers,
warped hands,
plastic skin,
excessive slow motion,
dreamy dissolves,
unreadable typography.

[おのふみ| AIクリエーター×個人開発 @onofumi_AI]

海外でバズってた「時間停止ワンカット」

Seedance 2.5で作った19秒（手持ちカメラで走る→転ぶ→時間が止まる→ドローンになって宙に浮いた小物を順番に舐める→時間が動き出す）。構造だけ借りて、中身を商店街のラーメン出前に。

条件はTopView経由のseedance2.5、480p、15秒。1本10.5クレジット、生成7分、修正なしの1発目がこれです。

驚いたのは日本語で、伝票の手書き「ラーメン 大盛 3つ 至急」と携帯画面の「店長 2:38」がそのまま読めた。店長との電話4往復も全部指定どおりに喋ってます。

プロンプトはリプに全文置きます。claudeに投げて人物と小物だけ差し替えれば、自分の商品でも同じ構造で作れます。

元ネタは広告に使ってたみたいですね。

[Ima Studio @ImaStudio_ai]

MiniMax H3 K-Fashion Ad | Cute Maximalism + High-Impact Editorial MV 💖CUTE IS BOLD. 💗

Created with @Hailuo_AI / MiniMax H3 × Ima Studio
#MiniMaxH3 #MiniMaxDesign

Prompt👇

Create a 15-second, 9:16, 24fps high-impact anime fashion commercial.
THEME:
SOLVÉRA — “KAWAII IS POWER”
STYLE:
Harajuku Kawaii × Decora × Y2K × Cyber Kawaii × Punk Kawaii × Japanese anime opening × premium fashion campaign × fast motion graphics.
Cute, sexy, wild, stylish, rebellious, youthful and visually explosive — never childish.
CHARACTER — HIGHEST PRIORITY:
One adult East Asian anime woman with refined Japanese-anime features, sharp playful eyes, slim feminine curves, long legs and confident fashion-model energy.
Keep her exact identity, face, hairstyle, age and body proportions consistent throughout.
Outfits may change, but never redesign the character.
VISUAL RULE:
Every shot should combine:
character motion + camera motion + foreground motion + background graphics.
Use strong parallax, foreground wipes, oversized hair clips, beads, hearts, stars, ribbons, plush charms, metallic bags, stickers, Purikura photos, manga speed lines, halftone, print misregistration, chromatic offset and huge typography.
Alternate camera moves:
Tracking / Orbit / Roll / Crane Up / Pull Back / Whip Pan / Foreground Pass.
0–1s — KAWAII HOOK
Extreme close-up through giant colorful accessories. 18mm camera rushes toward her face.
She lowers colorful sunglasses and looks directly at camera.
Huge text slams in:
“KAWAII!”
Small Japanese text:
“かわいい、爆発。”
RGB flash, sticker burst, giant heart wipes the lens.
1–2.2s — DECORA
Pink × yellow.
She walks toward camera while beads, ribbons, stars and clips fly onto her styling.
Low-angle tracking, slight camera roll.
Text:
“MORE! MORE!”
2.2–3.3s — HARAJUKU Y2K
Silver × cyan.
Metallic bag wipes across lens.
She appears, performs a fast turn and looks back over her shoulder.
Purikura frames and chrome graphics follow her movement.
Text:
“COLOR ME”
3.3–4.4s — SWEET LOLITA MIX
Lavender × mint.
Giant bows, lace and ribbons float around her.
Fast 90° orbit while she spins in the opposite direction.
Text:
“MIX IT UP”
4.4–5.6s — THREE-WAY SPLIT
Same heroine appears in three panels:
DECORA / Y2K / LOLITA.
Different camera motion and poses in each panel.
Large text across all three:
“SWITCH!”
5.6–6.8s — CYBER KAWAII
Black × cyan × hot pink × chrome.
She walks toward camera and reaches toward the lens.
Camera passes under her hand into a low angle.
Neon hearts and graphics orbit her.
Text:
“NO RULES”
Her hand covers the lens.
6.8–7.8s — TYPOGRAPHY EXPLOSION
Black × pink.
Huge Japanese typography:
“かわいいは自由。”
Repeated “NO RULES” layers in the background.
Camera flies through giant letterforms.
Torn paper, ink splashes, scratches, print-offset effects.
7.8–9s — PUNK KAWAII
Black × shocking pink.
She starts crouched, suddenly rises.
Low-angle camera cranes upward with slight roll.
She whips her cropped jacket as chains and bows explode outward.
Text:
“TOO MUCH?”
9–10s — MAXIMUM KAWAII
Extreme face close-up.
Rapid pull-back reveals a huge space overwhelmed by clips, stars, hearts, ribbons, plush charms and stickers.
She barely moves, only raises one eyebrow.
Text:
“まだ足りない。”
At 9.9s, EVERYTHING suddenly stops.
10–11.2s — SILENT BREAK
Pure white background.
Final LOOK: RAINBOW STREET.
More mature, refined and premium.
She slowly walks forward two steps while camera dollies backward.
Almost no VFX.
Only toy bell + two footsteps.
Pink ripple appears under her second step.
11.2–13s — VISUAL EXPLOSION RETURNS
Dozens of Purikura photo cards showing previous looks fill the space.
Camera snakes through a huge 3D collage of hearts, stars, ribbons, stickers and magazine graphics.
End on heroine in final look.
Text:
“HARAJUKU MODE”
13–15s — FINAL HERO
Hot pink background.
Heroine holds a confident, sexy fashion pose.
Huge white typography:
“KAWAII”
then
“IS POWER”
Then:
“かわいいは、強い。”
“SOLVÉRA”
“OUT NOW”
She tilts one shoulder and gives a confident smile.
Tiny heart pops.
Freeze final frame for 0.2s.
MUSIC:
Cute Hyperpop / Harajuku Kawaii Pop, around 150 BPM.
Heavy kick, clap, bubblegum bass, toy synth, sparkling bells, cute vocal chops.
0–5.6s: fast hyperpop
5.6–7.8s: darker cyber section
7.8–10s: maximum intensity
9.9s: complete stop
10–11.2s: bell + footsteps only
11.2s: full beat explodes back
13–15s: “KAWAII / IS POWER” hits sync exactly to the beat
AVOID:
childlike proportions, chibi style, schoolgirl look, face changes, body changes, identity drift, random outfit errors, static posing, weak camera motion, unreadable typography, generic cute visuals.
FINAL FEEL:
A high-budget Japanese anime fashion campaign combining anime-opening energy, fashion editorial styling and premium motion graphics — ultra-kawaii, sexy, rebellious, stylish and unforgettable.

Try MiniMax H3 on Ima Studio 👉
https://t.co/rxvRdUYoxo

#AIVideo #KFashion #MotionDesign #FashionFilm

[🟢DeepForest_Studio_AI🟢 @DeepForest_Inc]

MiniMax-H3 + ComfyUI-H3-Multishot を使用して引用元の65秒無言版から、今度は英語セリフ付きに挑戦！🙌

3枚のリファレンス画像を使って65秒の長時間生成✨
人物IDと背景の連続性を維持しながら、英語セリフもかなり安定。長尺でも最後までシームレスに生成出来た😎

#MiniMaxH3 #ComfyUI

補足😊

日本語のセリフも生成できますが、長尺で試していると、指定したセリフ以外の意味不明な発話が入ることが何度かありました。

そこで今回は、安定性を重視してセリフを英語に変更。指定した内容をかなり安定して話してくれました。少なくとも今回のテストでは、長尺の音声生成は日本語より英語の方が安定している印象です。

また今回は3枚のリファレンス画像を使用。動画内左側に表示している3枚が実際に使用したReference Imagesです。65秒を通して人物のIDはかなりよく維持され、廊下から図書館へ移動しても背景の連続性はほぼ保たれました。

一方で、長尺になるにつれて後半では口の動きと音声にほんの少しズレが出る場面もありました。また、人物のIDは維持されているものの、後半になるにつれて画質は少しずつ低下していきます。

それでも、人物・背景・音声を維持しながら65秒まで連続生成できたのはかなり面白い結果でした。まだまだ色々試せそうです😊

[すーさん @su_nagomi]

高級ファッション広告／実写エディトリアル／モーショングラフィックス／テロップアニメーション／フラッシュモーション

#SousakuAI のMinimax H3で動画作成
動画プロンプトはリプ欄に👇

[gane.wav @gane0wav]

MiniMax Designを使って、モーショングラフィック × MVの冒頭映像を制作してみました。

音楽とタイポグラフィ、映像を組み合わせたMV表現。
MVの幅が広がる✨

#MiniMaxH3 #MiniMaxDesign

[Heather Green @heathergreen]

#promptshare and sref for Midjourney
PROMPT: chubby orange cat sitting behind the counter of a tiny old bookstore, overflowing bookshelves, reading glasses, warm lamps, rainy evening outside the window, whimsical storybook illustration, highly detailed, cozy atmosphere --ar 5:6 --sref 3260656576 --hd

[Amira Zairi @azed_ai]

Prompt share: Timeless Line Elegance

💬 Prompt:
Elegant [subject] interacting with a [object], sophisticated black and white line art, delicate flowing outlines, refined illustration, minimal composition, graceful forms, luxury editorial aesthetic, crisp ink strokes, clean white background

Try it and share yours ✨

[蛋黄堡 @Hamburgerai]

【中式庭院】转绘成水墨风格

建筑没有被一笔一笔描满。
一整块浓墨压住屋顶，柱廊、台阶和树影只留下必要的线索。
照片里的园子到了纸上，忽然安静了很多。
颇具东方韵味

prompt：
请将我上传的每一张照片分别制作成一张独立的高级建筑观察海报，不进行多图拼贴，每张照片单独输出。

整体优先采用4:3横版左右双联构图。左侧约占画面50%，忠实保留原始照片；右侧约占画面50%，将同一场景重新诠释为“墨线设色建筑观察画”。

左右区域的中间边界必须笔直、清晰、利落地硬切。不要渐隐、羽化、撕纸、雾化或相互渗透，也不要让照片、墨线、植物或色块跨越边界。边界本身不需要额外描线。

左侧保留原始照片的主体身份、建筑结构、地形关系、空间方向、自然光影、真实质感和原有色彩氛围，仅进行轻微、克制的高级摄影调色，使其具有建筑杂志、博物馆出版物和艺术展览摄影的质感。不得拉伸、扭曲、移动、替换或重新设计主体。

右侧以温暖的米白色天然纸张为背景，观察并提取照片中最具识别度的建筑母题，例如屋顶、飞檐、门洞、柱列、平台、墙体、台阶、栏杆、廊桥以及主要空间方向，将其重新组织为现代东方纸本建筑观察画。

不要完整复刻照片，也不要把建筑的所有细节都画出来。用一至两块浓黑或灰黑色的不规则墨形建立最主要的建筑轮廓和视觉重量，例如用一整片浓墨概括屋顶、檐口或墙体；再用少量纤细、略有断裂的灰黑墨线提示柱列、栏杆、透视方向和必要的结构节点。

遵循：

整体建筑主形 > 空间方向 > 必要结构 > 微小细节。

瓦片、窗格、砖纹、雕花、斗拱和栏杆不得逐一精确描绘。复杂结构应主动删除、合并和概括，使人能够一眼辨认原建筑，同时保留手工观察、现场速写和艺术化删减的痕迹。

植物不要逐叶刻画。将树冠、竹林、灌木和远处植被压缩为少量灰绿、苔藓绿、鼠尾草绿、橄榄绿或赭色的不规则点染与干刷色块。树干和枝条只用极少量深色线条提示。

如果照片中包含水面，仅使用几组断续、稀疏的灰蓝或灰绿色水平笔触表现，不要完整填满水域。天空、空气、地面、雾气和距离可以直接由未经绘制的纸张留白代表。

右侧至少保留40%至60%的连续纸张留白。所有墨形、线条和设色应围绕一个主要建筑母题集中成组，不要将零碎笔触平均散布到整张纸面，也不要为了交代完整环境而补满背景。

配色从原始照片中提取，但压缩为克制的有限色盘。以墨黑、灰黑、暖灰、灰绿、苔藓绿、矿物蓝绿、赭石和米白色为主，仅保留一处小面积暖色作为视觉焦点。

如果原照片中存在重要人物，可以将其保留为唯一的小面积朱红色人物剪影；如果没有人物，不得凭空添加。照片中具有识别意义的红柱、灯笼、门扇或标志性构件，也可以作为唯一的朱红色焦点，但不要让多个暖色元素争夺注意力。

使用真实纸本绘画质感：

bold black ink mass, broken dry brush, sparse structural linework, muted mineral color, natural ink absorption, rough brush edge, subtle pigment granulation, visible paper fibers, warm ivory paper, partially exposed paper.

墨线应具有轻微迟疑、干涩、断裂和手工误差。设色应少量、不透明或半不透明，避免大面积透明水彩晕染。画面是现代建筑观察、东方留白意识与纸本编辑设计的结合，不是完整传统水墨山水。

可在右侧留白区域加入极少量排版信息：根据场景提炼2至4个英文单词作为标题，并搭配编号、地点或年份。使用细瘦、克制、具有建筑出版物气质的衬线字体或无衬线字体。文字不得压住建筑主形，不要加入虚构标志、印章、签名或书法题字。

整体气质应当安静、克制、清晰、理性而富有诗意，像建筑师旅行途中完成的一页观察记录，也像现代美术馆出版的东方建筑研究海报。

Style keywords:

ink and color architecture observation, East Asian architectural field drawing, bold black roof mass, sparse structural linework, muted mineral pigments, warm ivory paper, controlled negative space, dry brush texture, architectural observation sketch, modern Asian editorial poster, museum architecture monograph, clear split boundary.

Avoid:

blended photo-paper boundary, feathered transition, elements crossing the center boundary, busy perspective web, complete building rendering, exact roof tiles, detailed windows, individual leaves, dense foliage, photorealistic illustration, dominant watercolor wash, traditional full landscape painting, CAD blueprint, clean vector geometry, smooth digital gradients, generic ink filter, fake calligraphy, signature, seal, logo, watermark, cluttered typography.

最重要的规则：右侧只用一个具有重量感的浓墨建筑主形统领画面，其余结构线、植物和设色只负责必要提示。清晰边界、连续留白与主动删减，与建筑本身同样重要。

[ア @yuruyurau]

a=(m,d=mag(k=9*cos(i*5)*sin(i),e=cos(i*3)*cos(i*2)*9)**3/1999+1.5-sin(t/2+m)**3/3)=>point(99*sin(c=d/16-t/48+m)+k*(p=d**sin(d*d-t+m))+200,99*sin(c*4)+e*p+200)
t=0,draw=$=>{t||createCanvas(w=400,w);background(9).stroke(w,96);for(t+=PI/20,i=1e4;i--;)a(i%16*13)}//#つぶやきProcessing

[Michael Rabone @michaelrabone]

Tribal sunset & motion

Arikara chief with white tribal face paint, buffalo and deerskin garments, feathers and jewellery, cinematic, golden hour, bokeh, dust --ar 16:9 --raw --sref 1817778782::4 3349256275::3 2782365120::2 2645245632::2 2594221865 --sw 100 --s 300 --hd --v 8.2

[simeon-sanai @Naiknelofar788]

GPT image 2

Prompt:
Create a premium vertical 4:5 travel poster of [CITY], [COUNTRY], designed entirely as a handcrafted layered 3D paper-cut miniature world.

Use a soft pastel paper background with subtle paper grain, sculpted paper clouds, and multiple layers of textured cardstock. Build a beautiful miniature cityscape with realistic cut-paper edges, tactile paper textures, gentle shadows, raised 3D elements, and elegant handcrafted depth.

Make [MAIN ICONIC LANDMARK] the dominant central feature, surrounded by 4–6 recognizable landmarks, architecture, streets, bridges, nature, cultural elements, and local details that instantly represent the city. Add a river, lake, coastline, or other important geographical feature where appropriate.

Include tiny paper-cut people, vehicles, boats, trees, cafés, or street details to create scale and storytelling while keeping the composition clean and sophisticated.

Color palette: harmonious soft pastels inspired by the city and country, with cream, muted green, warm beige, dusty pink, soft blue, terracotta, and complementary tones.

At the top, add elegant minimal typography:

[CITY]
[COUNTRY]
[SHORT ICONIC TAGLINE]

Premium editorial travel-poster composition, whimsical handcrafted paper sculpture, realistic architectural proportions, layered depth, delicate paper fibers, soft studio lighting, subtle ambient shadows, sophisticated composition, highly detailed, charming miniature world, clean negative space, artistic and collectible poster aesthetic.

Avoid: photorealism, plastic, glossy CGI, metallic surfaces, clutter, distorted landmarks, excessive text, oversaturated colors.

Keep the same visual structure and artistic style across different cities, while changing the landmarks, cultural details, geography, and color accents to represent each destination authentically.

[노노아카이브 @nono_ai_archive]

심심한 이미지에
GPT로 텍스쳐만 입혀보세요.

1.제록스 텍스쳐 (Xerox)
2.하프톤 (Halftone)
3.디스트레스 (Destressed)

[Sairah @Sairah_0]

GPT Image 2 On ChatGPT

Prompt:

Use the uploaded photo as the exact background reference. Preserve the original real-life environment, architecture, perspective, lighting, colors, shadows, and camera composition.

Add a cute, stylish black-and-white hand-drawn anime doodle girl naturally into the real photograph. The character should look like a 2D ink illustration placed physically inside the real-world scene, creating a realistic live-action photo + hand-drawn illustration fusion.

Character style:
- Clean black ink line art
- White/very light interior with minimal shading
- Cute Korean/Japanese manga-inspired aesthetic
- Natural, elegant proportions
- Long flowing dark hair
- Simple facial features with expressive eyes
- Clearly visible black outline around the entire character, like a hand-cut sticker
- Slightly imperfect hand-drawn lines for an authentic sketchbook feel
- No colors on the character except black and white

Character integration:
- Match the character's scale, perspective, and position to the photograph
- Make her feet properly contact the ground/stairs
- Add realistic interaction with the environment through natural shadows and overlapping perspective
- Keep the background completely photographic and detailed
- Do not turn the background into anime or illustration
- The character should appear as if a hand-drawn person has stepped into the real photograph

For the outfit, use casual modern streetwear appropriate for the location. Preserve recognizable clothing details such as oversized T-shirts, jeans, sneakers, shoulder bag, jersey numbers, or accessories when specified in the reference image.

Composition:
- Full-body character when possible
- Character positioned naturally in the foreground or middle ground
- Keep the original landmark, stadium, buildings, stairs, trees, sky, advertisements, and surroundings clearly visible
- Maintain the original photograph's wide-angle composition
- Bright natural daylight
- Crisp details
- Realistic photographic background
- Strong contrast between the monochrome illustrated character and colorful real environment

Overall aesthetic: cinematic travel photography + Korean street photography + monochrome manga line art + scrapbook/sticker illustration, with seamless perspective and lighting integration.

Do not make the entire image cartoon-like. Only the character should be illustrated; everything else must remain photorealistic.

[Kiki @Mayz1169]

MiniMax H3 is really fun for creating anime PVs.

The character transformations, graphic effects, and fast-paced editing are exactly the kind of style I love.

Prompt below ⤵️

#MiniMaxH3 #MiniMaxDesign