十豆加

[老杨啊 @yhslgg]

https://t.co/yud2z8fiNM

[鸟哥 | 蓝鸟会🕊️ @NFTCPS]

兄弟们，AI 拍视频卷到这种程度了？？

刚刷到这个 Seedance2-skill，直接给我看傻了 👇

🔗 https://t.co/pwPSegz3FC

1️⃣ 电影级提示词外挂，专门喂字节 Seedance 2.0（即梦）
2️⃣ 教 AI 写镜头语言、分镜结构、@素材引用，全是专业那一套
3️⃣ 塞进 Cursor、Claude Code 里直接用，一键出片

以前写 prompt 靠玄学，现在直接抄作业，普通人也能拍出运镜大片，这波是真的要把视频制作门槛踩到地板了。

[𝗦𝗮𝗻𝗶𝗮 @saniaspeaks_]

GPT image 2 on Chatgpt

Prompt:

A Korean young woman with the same unchanged face, pale soft complexion blown slightly by harsh 90s disposable-camera flash. Loose black hair falling forward as she leans toward a scratched bathroom mirror. Subtle knowing smirk. She’s wearing a soft faded peach tank top and loose cotton lounge pants. The mirror edges are chipped, toothpaste and skincare clutter visible on the sink. Flash reflection creates glare and lens flare. Grainy, raw, imperfect framing, authentic late-night candid energy.

[ANKIT PATEL 🇮🇳 | AI @Ankit_patel211]

GPT Image 2 ChatGPT

Ultra-realistic 16:9 close-up studio photograph of a French citizenship identity card placed on a clean light grey or white background. The card is horizontally aligned, perfectly centered, captured in sharp focus with soft diffused studio lighting.
The card design follows an official European government ID aesthetic with refined typography, balanced layout, and advanced anti-counterfeit detailing. The top header reads: “RÉPUBLIQUE FRANÇAISE” with a smaller subtitle “CARTE DE CITOYENNETÉ / CITIZENSHIP CARD.”
Top-left corner includes the French Marianne emblem with the national motto “LIBERTÉ · ÉGALITÉ · FRATERNITÉ.” Top-right features a European-style blue rectangle with yellow stars and country code “FR.”
The card contains a passport-style portrait of a young adult French woman:
– Natural light skin tone with subtle freckles
– Medium-length soft brown hair, slightly wavy, center-parted
– Hazel or light brown eyes, neutral expression, direct gaze
– Minimal makeup, realistic skin texture
– Wearing a simple dark top and small gold stud earrings
Personal identity details displayed clearly in bilingual format (French / English):
– Nom / Surname: DUPONT
– Prénoms / Given Names: LÉA MARIE
– Date de naissance / Date of Birth: 14 JUIL / JUL 1996
– Lieu de naissance / Place of Birth: LYON (69), FRANCE
– Nationalité / Nationality: FRANÇAISE
– N° de citoyenneté / Citizenship No.: FR-960714-123456
– Date de délivrance / Date of Issue: 22 MAI / MAY 2024
– Date d’expiration / Date of Expiry: 22 MAI / MAY 2034
– Taille / Height: 1,65 m
Security and design features include:
– Intricate guilloche patterns across entire card surface
– Fine-line engraving textures similar to banknote printing
– Microprinted text layers and anti-counterfeit linework
– Holographic security patch with iridescent reflections
– Circular “RF” (République Française) seal hologram
– Subtle engraved illustration of Marianne profile on right side
– Architectural engraving of a classical French building in the background
– Tricolor accent (blue, white, red) flowing along the bottom edge
– Holder’s signature at bottom right: “Léa M. Dupont”
Material appearance: laminated polycarbonate card with slight gloss, visible light reflections, and layered holographic overlays.
Lighting is soft and even, emphasizing texture, print detail, and holographic elements without harsh shadows. No hand visible. Clean, minimal, official composition.
Style: ultra-detailed, photorealistic, 8K resolution, macro clarity, professional product photography, documentary realism.

[underwood @underwoodxie96]

love this prompt！
prompt：https://t.co/ObtMCtVau1
Please create an approximately 4-page swimsuit fashion magazine featuring a live-action cosplay of 【Name】 as the main character. Outfits & Styling: Have her wear a variety of swimsuits in different colors, patterns, and shapes, and give her different hairstyles as well. Expressions & Poses: Use a wide range of expressions, such as shy, seductively biting her lip, confident, and a longing expression with her tongue out. The poses should also be intentionally varied to avoid repetition. Layout Elements: Please include magazine-style catchphrases (catchcopy) and titles. A personal profile section is not needed. Visual Style: Effects Control: Do not overuse glitter, grain, or scattered light effects. Only use them sparingly in areas where they are truly needed. Image Stability: Be careful not to let the image fall apart due to overly detailed rendering (such as distorted limbs or structural issues). Maintain a high level of detail while ensuring the overall image remains harmonious and well-balanced.

[Ciri @Ciri_ai]

GPT Image 2 on Chatgpt

Prompt: Create an illustration of how you imagine “me”:

Korean-style illustration, with a distinct and memorable character design, natural and expressive facial emotions, half-body composition, and a dynamic pose. The clothing should feature intricate, refined details. Use a hand-drawn doodle style with ink splashes, loose and spontaneous linework, and a mix of pastel tones with ink shading. Aim for a comic sketch texture. Keep the background minimal and white, with symbolic decorative elements scattered around the character. The overall atmosphere should be strong and evocative, with high detail and high quality.

[DStudioproject @D_studioproject]

How to make ramen with GPT Image 2 x Seedance 2.0 Anime Style on higgsfield

workflow⬇️

[雨哥向前冲 @xiangxiang103]

AI 虚拟女友一个月赚4.3万美金！

我今天看到这个帖子，直接看呆了。

一个21岁的 Austin 大学生，用 AI 造了一个叫 Maya 的 OnlyFans 账号，第一个月就赚了4.3万美元。 问题是——照片里那个女孩，根本不存在。

Maya 22岁，UCF 心理学辍学生，1247个付费订阅者，单月最高一个粉丝就刷了1847美元。

她没有拍视频，没有真人打字，所有消息都是 Claude Code 写的，所有照片是 Flux 生成的，声音是 ElevenLabs 合成的。整个 Maya 只是他 MacBook 里四个 .md 文件。

以前造一个 AI 虚拟网红要花半年甚至一年半，现在只需要4个星期，下一个可能就是一个周末。

作者详细写了 Maya 的四个文件：
persona.md 写她完整的人设、背景、喜好；
voice.md 锁声音风格；
flux.md 固定长相和拍摄风格；
brain.md 记录每个订阅者的记忆和互动历史。
Claude Code 每次回复前都会读这四个文件，从不崩人设，也从不忘记粉丝说过什么。

真的很感慨，技术栈已经简单到可怕的地步。

以前的 Aitana López 花了18个月，现在 Maya 4周就跑通了。

这已经不是科幻，而是任何一个有 laptop 的人都能复制的路径——从 OnlyFans 到 Instagram、TikTok、Twitch，随便哪个平台都能套。

AI 让“虚拟人”成本降到了几乎为零，剩下的唯一瓶颈，是你有没有那个 taste 去决定这个虚拟人要成为什么样子。

[MapleShaw @msjiaozhu]

还记得之前火遍全网的来自神佬 @berryxia 的角色拆解提示词不？

搬到 GPT Image 2 试了下，质感比香蕉上了一个档次！😍

提示词👇
基于用户提供的参考图片，创作一张超高品质、照片级写实的3D角色拆解艺术海报。将人物转换为Unreal Engine 5 (Metahuman) 级别的超写实数字人，并将其极具个人魅力的私人物品以严谨的"Knolling"风格进行布局展示。

画面比例：16:9 横版
艺术风格核心：时尚超写实主义 (Fashion Hyper-realism) — 极致的皮肤质感、奢华的材质表现、诱惑的电影氛围。
质量标杆：对标 Tom Ford/Gucci 的时尚广告大片、赛博朋克2077的高端角色建模、Vogue 摄影作品。

📷 物品布局 (Item Layout) - Knolling 放射式构图
总物品数：25-30件，围绕角色呈放射状排列，注重留白与呼吸感。

分类1：高定穿搭 (Haute Couture) - 深红丝绒标签
- 主服装拆解：
- 如果是女性：黑色丝绸吊带裙（展现丝滑垂坠感）、极细高跟鞋（红底）、蕾丝内衬细节、透明黑丝袜（展现透肤度和光泽）。
- 如果是男性：解开领口的丝质衬衫、定制西装马甲、手工皮鞋（镜面光泽）。
- 材质重点：重点表现丝绸的流光、皮革的纹理、肌肤与布料接触的压痕。

分类2：私密美妆 (Intimate Beauty) - 香槟金标签
- 妆容细节：口红（膏体带水润光泽，管身留有指纹）、眼影盘（碎钻闪粉质感）、昂贵的复古香水瓶（琥珀色液体）。
- 氛围道具：一支燃烧了一半的香薰蜡烛（带烟雾）、一杯红酒（挂杯细节清晰）。
*示例：口红盖子打开，旁边有一张印着吻痕的餐巾纸。*

分类3：奢华配饰 (Luxury Accessories) - 铂金标签
- 珠宝：钻石项链（强烈的色散火彩）、蛇形手镯、机械腕表（透视机芯）。
- 随身物：真皮手包（细腻的荔枝纹）、金属烟盒或雪茄剪、豪华跑车钥匙。

分类4：欲望符号 (Symbols of Desire) - 霓虹紫标签
- 道具：一副复古墨镜、胶片相机、一本未读完的时尚杂志、黑胶唱片。
- 暗示性物品：散落的珍珠、一朵红玫瑰（花瓣带露珠）。

📷 角色呈现 (Character Presentation)
- 姿态：不再是标准的站立，而是略带慵懒、自信的姿势（如坐在隐形的高脚椅上，眼神直视镜头，带有强烈的吸引力）。
- 皮肤细节：必须展现真实的毛孔、细微的汗毛、嘴唇的纹理、锁骨的阴影。皮肤要有“微出汗”的光泽感(Dewy Skin)。

📷 爆炸视图技术 (Exploded View Technique)
- 连接线：极细的金色激光线，带有极简的科技感。
- 引导设计：不再使用普通箭头，而是使用光晕引导视线。
- 技术注释：字体极小且锐利，类似于高端医学美容或时尚杂志的分析图。

📷 视觉包装 · THE AESTHETICS OF ALLURE 📷
- 字体：Didot 或 Bodoni（经典的时尚衬线体），极细字重，白色或金色。
- 副标题 (手写体)："Elegance is an attitude / 优雅即是态度"
- 色彩方案：
- 深邃暗调：午夜蓝(#000033) + 勃艮第红(#800020) + 香槟金(#D4AF37)。
- 赛博魅惑：黑色背景 + 霓虹紫(#B026FF) + 电光蓝(#00FFFF)。

📷 背景与氛围 (Background & Atmosphere)
- 光影设计 (Lighting)：
- **伦勃朗光 (Rembrandt Lighting)**：在面部形成戏剧性的三角光区。
- **轮廓光 (Rim Light)**：强烈的背光勾勒出角色的身体曲线和发丝边缘。
- 背景：深色背景，带有焦散光斑(Caustics)或模糊的城市夜景灯光(City Bokeh)。
- 特效：轻微的烟雾缭绕，增加神秘感和空气感。

📷 技术规格 (Technical Specifications)
- 渲染引擎：Octane Render 或 Unreal Engine 5。
- 物理材质 (PBR)：
- **皮肤**：真实的人类皮肤着色器，包含次表面散射(SSS)和微量油脂层。
- **布料**：丝绸必须有各向异性(Anisotropic)反光；丝袜要有半透明Alpha通道处理。
- **金属**：高反射，带指纹污渍的真实感。
- 摄影机：85mm 人像焦段，大光圈 (f/1.8) 虚化背景，焦点锐利地锁定在角色眼睛和关键物品上。

📷 后期处理 (Post-Processing)
- 色调：电影感调色 (Teal & Orange 或 Black & Gold)。
- 辉光：高光处添加明显的柔光滤镜效果 (Pro-Mist Filter look)，让皮肤看起来更梦幻。
- 锐化：强化眼部和珠宝的锐度。

📷 质量基准：
画面必须摆脱“塑料感”和“模型感”，达到甚至超越 8K 摄影照片的质感。角色必须具有灵魂和性张力，眼神能穿透屏幕。

[serein ｜买美股上币安 @you1873118]

GPT-Image-2的一个妙用形象

这个提示词是一个老哥发我的

我试了下，效果也太强了

如果放在一年前

你可能三天三夜的ComfyUI工作流

才能做出来

提示词：

请根据我上传的人像照片，做一套个人形象分析图卡，包含发型、妆容、色彩和珠宝。要求：保留五官脸型肤色，不要过度修图，所有变化在同一张脸真实展示，风格干净高级。发型：长短、卷直、刘海，对比最适合/ 普通/不建议（显脸小、显老）妆容：眉眼鼻唇分析，标签（自然、提气色、柔和）色彩：不同颜色上身，对比推荐/普通/不适合（显白、显老）珠宝：珍珠、翡翠、红蓝宝、钻石、黄金，对比推荐/普通/不建议 整体：视觉为主，文字简短，4:5比例

[Berryxia.AI @berryxia]

比起疯传的那个风格，我自己更偏爱这个风格。

Prompt：
「彩色潦草小狗线条风格绘制该图，童趣和doodle加入其中，务必使用毫无章法的绘制手法，凌乱和草率即可。』

[Larus Canus @MrLarus]

太飒了！
🤯用ChatGPT+Seedance生成街头篮球1v1，真实感很强！

运球试探、crossover变向、欧洲步拉杆反手上篮⛹
GPT先生成16宫格动作图，再生成视频
爆发力、节奏感和对抗感拉满！

提示词放在评论区👇

[FFEE知识分享 @FFEE_2025]

换脸工具🛠️

提供3个本地AI换脸工具，请合理使用哦

链接：https://t.co/p54KbdQADi

[Rossy @xRahultripathi]

Just giving my self a art off peace with
Trend

Created image using GPT 2.0 Created on @renoiseai

Prompt ⤵️

Chibi Avatar Scrapbook Photo Prompt (Based on Uploaded Female Portrait)

Use the uploaded image of a young woman as the base. Do NOT add any new person. Transform the existing subject into a chibi scrapbook-style composition while fully preserving her real identity, face, hairstyle, hair color (long soft brown hair), natural makeup, black outfit, pose, lighting, and cozy indoor aesthetic.

[Theme Detection]
The scene reflects a cozy lifestyle / selfie / workspace vibe with soft lighting, aesthetic room elements, a motivational neon sign, and a laptop desk setup. Maintain a calm, stylish, slightly productive lifestyle mood throughout.

[Chibi Avatars]
Add 5–8 chibi mini versions of the SAME girl:
- same long soft brown hair (slight waves)
- same outfit tone (black top)
- same gentle, confident expression

Style:
- kawaii chibi sticker style
- big head, small body proportions
- expressive cute eyes
- soft pastel shading
- clean, high-quality sticker finish

[Chibi Actions – no repetition]
- holding coffee
- working on laptop
- taking selfie
- relaxed leaning pose
- peace sign
- thinking pose
- stretching / chill pose
- smiling & waving

[Sticker Placement]
- Place chibi avatars around edges and empty areas only
- Do NOT cover the main face or body
- Add white outline + soft drop shadow
- Slight floating sticker effect

[Doodles]
Add subtle scrapbook-style doodles:
- hearts, sparkles, stars, arrows
- aesthetic lines and circles
- soft glowing accents

Color palette:
white + warm yellow + soft pink (matching neon vibe)

[Handwritten Text]
Add 5–8 short aesthetic phrases:
- "just me "
- "soft life"
- "stay cozy"
- "little moments"
- "focus mode"
- "good vibes only"
- "create & grow"
- "be you "

Style:
- handwritten font
- white text with soft glow
- selective highlights in pink/yellow

[Composition]
- Center focus: real girl (unchanged)
- Surround with chibi stickers + doodles + text
- Keep layout balanced, clean, not cluttered

[Style Keywords]
realistic photo + chibi stickers, kawaii aesthetic, cozy lifestyle, warm lighting, scrapbook diary, soft glow, high detail, Instagram aesthetic

[Final Output]
A polished, high-quality, social-media-ready scrapbook image combining:
realistic subject + cute chibi extensions + aesthetic doodles

[Avoid]
- altering facial identity
- over-beautifying or AI face distortion
- cluttered composition
- unrelated/random elements
- covering the subject

[underwood @underwoodxie96]

GPT Image 2.0
https://t.co/geBdeEx36Q
prompt:
{ "subject": { "description": "A young adult woman taking a close bathroom mirror selfie, wearing tall black bunny ears, creating a playful yet private contrast, with the low-resolution realism of a casual phone snapshot", "mirror_rules": "Mirror selfie perspective, facing the mirror and phone, slight compression and softness, preserving the layered reflection of the bathroom mirror", "age": "young adult woman, approximately in her 20s", "expression": { "eyes": { "look": "half-open, relaxed, slightly dazed", "energy": "lazy, intimate, as if she has just spoken or finished a soft phrase", "direction": "looking toward the camera direction, gaze slightly lowered and not fully focused" }, "mouth": { "position": "lips slightly parted, a hint of teeth visible, as if breathing or speaking quietly", "energy": "natural, unforced, subtly playful" }, "overall": "relaxed expression with a private late-night selfie feeling and a faintly unfocused mood" }, "face": { "preserve_original": "Preserve the soft oval face, slim nose bridge, natural lip shape, and slight selfie-angle asymmetry; avoid excessive retouching", "makeup": "Inspired by East Asian casual selfie beauty aesthetics: translucent soft-matte base, subtle under-eye softness, restrained eyeliner, naturally rosy glossy lips, like everyday makeup under warm indoor bathroom light" } }, "hair": { "color": "deep black to natural black-brown", "style": "long straight hair with volume, bangs and side strands falling naturally, one side partially covering the cheek and eye area", "effect": "hair has localized shine from the bathroom ceiling light, slightly messy and fluffy, not commercially styled, as if quickly adjusted" }, "body": { "frame": "close upper-body framing, slim shoulder and neck line, body centered slightly lower in the frame", "waist": "waist not visible, do not emphasize", "chest": "upper chest and collarbone area partially visible, main chest covered by clothing, keeping a natural upper-body silhouette", "legs": "legs not visible", "skin": { "visible_areas": "visible face, neck, collarbones, both shoulders, upper chest edge, front of both arms, hands and fingers; waist, legs, back, and hips are not visible", "tone": "fair warm-toned skin, appearing slightly peachy under the warm bathroom ceiling light", "texture": "skin appears soft and fine but not plastic-smooth; shoulders, neck, and hands feel visually warm and natural to the touch, with subtle knuckle shadows and real skin variation preserved", "lighting_effect": "top light falls from above and slightly forward, creating warm highlights on the forehead, nose bridge, shoulders, and collarbones; hair casts shadow over one side of the face, with slight overexposure and phone compression along skin edges" } }, "pose": { "position": "facing the mirror, head slightly leaning forward, face close to the camera", "base": "both hands raised, holding or adjusting the shoulder straps and front fabric near the upper chest", "overall": "casual, close, slightly unstable pose, like a spontaneous shot taken while adjusting clothing in front of the mirror" }, "clothing": { "top": { "type": "black spaghetti-strap camisole or lace-trim top", "color": "black body with white lace or pale decorative trim on the straps", "details": "dark lace texture at the front, narrow straps, fabric fitting the upper body", "effect": "fabric appears soft and matte in low light, creating strong contrast against fair skin" }, "bottom": { "type": "not visible", "color": "not visible", "details": "do not describe" } }, "accessories": { "headwear": "tall black bunny-ear headband, extending out of the top frame, with visible white lace or bow detail on one side", "jewelry": "no obvious jewelry visible", "device": "phone positioned low in front of the mirror, with a dark phone edge or obstruction visible at the bottom of the frame", "prop": "no additional handheld prop" }, "photography": { "camera_style": "low-resolution phone mirror selfie, authentic social-media snapshot feel, slight noise and compression artifacts", "angle": "camera slightly below face level or near chest height, subtly looking upward, placing the face near the center while the bunny ears extend vertically", "shot_type": "close half-body selfie, head, shoulders, and upper chest framing", "aspect_ratio": "vertical, close to a 9:16 phone screenshot ratio", "texture": "soft focus, slight blur, warm white balance, localized overexposure, like an instant photo under bathroom lighting", "lighting": "warm recessed bathroom ceiling light directly from above, creating highlights on the forehead and shoulders; tiled background reflects soft ambient light, while hair shadow darkens one side of the face", "depth_of_field": "phone small-sensor deep focus overall, but close selfie distance and compression make the face and hair edges slightly soft" }, "background": { "setting": "indoor bathroom or washroom mirror area", "wall_color": "beige, light brown-gray tiled wall", "elements": [ "bathroom tiles", "warm ceiling downlight", "towel rack or shelf on the right", "light-colored towel and green object", "mirror edge and reflected bathroom space" ], "atmosphere": "private, late-night, spontaneous, slightly messy but realistic home bathroom mood", "lighting": "mainly warm indoor ceiling light, soft background reflection, low-light feel with mild phone noise" }, "the_vibe": { "energy": "lazy, close, playful, slightly dazed", "mood": "like a paused moment after getting ready, leaving behind the emotional residue of an unfinished gesture", "aesthetic": "casual bathroom selfie, bunny-ear accessory, low-resolution social-media photo, subtle nighttime mood", "authenticity": "preserve imperfect framing, hair covering part of the face, phone compression, hard ceiling highlights, and background clutter", "intimacy": "close mirror selfie intimacy, like a casual photo meant for someone familiar", "story": "She may have just been adjusting her straps, looked up at the mirror, and pressed the shutter without overthinking", "caption_energy": "“Almost ready, but not really ready yet.”" }, "constraints": { "must_keep": [ "black bunny-ear headband", "bathroom mirror selfie", "close vertical framing", "warm ceiling light", "hair covering one side of the face", "naturally parted lips", "black spaghetti-strap top with white lace trim", "both hands holding the straps or front fabric", "authentic low-resolution phone texture" ], "avoid": [ "overly polished studio look", "plastic skin smoothing", "excessively sexualized body emphasis", "luxury spotless bathroom", "cinematic controlled lighting", "turning it into a full-body shot", "cartoonish bunny ears", "over-retouched influencer face" ] }, "negative_prompt": [ "over polished", "plastic skin", "studio lighting", "luxury bathroom", "perfect symmetry", "oversexualized pose", "wide angle distortion", "cartoon bunny ears", "heavy makeup", "sharp commercial photo", "extra fingers", "deformed hands", "wrong mirror reflection" ] }

[币圈女菩萨 | Pizza披萨🍕 @xincctnnq]

codex 生成图片真的很厉害。
这已经肉眼无法辨别真假了吧。

我只是在网上找了几张自己喜欢的风格，
然后把我的脸的正面侧面照发给了 codex。

下面就是它给我生成的图片。
这样下去我感觉我可以去 instagram 运营一些新业务了……
lol 🤣

这年头，照都不用自己拍了。

[大卫叔的AI旅程 @aidavid125]

高端人物杂志封面.
---
Prompt:

High-end fashion magazine cover, title "YOU", portrait of a mysterious elegant woman, black background with soft gradient light,

wearing structured couture outfit, satin and silk textures, minimal exposure, luxury styling,

dramatic cinematic lighting, high contrast, deep shadows, premium editorial photography,

black and gold color palette, ultra refined, fashion week aesthetic,

8K, sharp focus, magazine layout typography, bold serif masthead --ar 2:3 --style raw
---

[MapleShaw @msjiaozhu]

卧槽！我必须单方面宣布 Codex 已经成为我的挚爱！😍

今天早上打开电脑，发现 Codex 昨晚自己给我跑了一批图。

🥳有几张真的开始有那种“相册里翻到的生活切片”的感觉了：糊的、歪的、手抖的、构图不完美，但情绪是对的。

我本来只是想试一下在 Codex 批量生图，结果现在感觉快被我养成一个会自动探索的个人 Prompt Agent 了！

[Latte @0xbisc]

Kitchen Hunt

GPT Image 2 + Seedance 2.0 is the best combo right now for speeding up creative workflows.

Prompt below ↓

[AmirMušić @AmirMushich]

Ok guys, there you go with the prompts:

Redraw the attached image in the most clumsy, scribbly, and utterly pathetic way possible. Use a white background, and make it look like it was drawn in MS Paint with a mouse. It should be vaguely similar but also not really, kind of matching but also off in a confusing, awkward way, with that low-quality pixel-by-pixel feel that really emphasizes how ridiculously bad it is. Actually, you know what, whatever, just draw it however you want.