Get live statistics and analysis of 古一's profile on X / Twitter

继续复刻！ 这次挑选了一张karin点赞量最高的作品，复刻难度也不小，炼了一晚上😅。 主要难点在于整体氛围感的拿捏和女生的妆容，拆解下提示词重点（完整prompt放评论区，记得回来交作业）： 1.女生的妆造描述 一开始腮红老是打太重，跟年画娃娃似的，后面加了一句好多了：faint pink flush on the cheeks, soft and natural lighting 2.物理环境描述 居酒屋，被中年大叔环绕，货架等描述用以复现画面中的人群结构和社会氛围。 3.视觉风格描述 主要强调低对比度，复古氛围，35mm胶片的细腻颗粒感，给了一些风格参考森山大道和王家卫。 最后精选了4张，你更爱哪张？

完美复刻karin，可以起号了！😎 试了很多个模型seeddream，banana，recraft，只有midjourney才能找到这种日系朦胧感的精髓。 提示词如下，大家可以去试试： a dreamy intimate portrait of a beautiful realistic asian woman lying under a blanket, soft natural light, cozy warm bedroom, cinematic film grain, 35mm film aesthetic, shallow depth of field, soft focus, delicate skin texture, natural makeup, slightly messy hair, Fujifilm Pro 400H look, pastel warm tones, realistic lighting, subtle haze, gentle atmosphere, by Petra Collins and Ren Hang --ar 3:4 --v7 --style raw

换一种摄影风格，OK吗？ 咒语： A captivating black and white close-up portrait of a beautiful young Asian woman with flawless skin, short bob haircut featuring straight blunt bangs framing her forehead, eyes gently closed with long lashes, subtle pouty lips, serene and introspective expression, wearing a form-fitting deep V-neck black ribbed sweater that accentuates her, a delicate thin necklace with a small pendant, dramatic side lighting from the left creating strong shadows and highlights on her face, neck, and shoulders, high contrast chiaroscuro style, minimalist studio background with soft gradient from light to dark, photorealistic, vintage film grain effect, high detail, emotional and mysterious atmosphere.

#前女友来了 咒语： A stealthy candid capture of a Asian beauty at a night market, straight hair with bangs, rosy cheeks, wearing a deep-neck tank top and jeans, browsing stalls with a curious gaze, low-light with neon signs blurring in the background, grainy texture like from a phone in dim conditions.

这俩妹子身材巨好，跳得也很好看 唯一不足的就是： 每个舞蹈视频选的bgm很不搭 当舞蹈动作和音乐节奏高度同步时，我们的大脑会产生一种强烈的预期满足和时间对齐的快感 相反，如果不同步，大脑的预期被打破，就会带来不适感 视频来源：视频号-C颖Emliy

想用AI出动漫图，只知道吉卜力？ 那你至少错过了 90% 的宝藏风格！ 大友克洋、伊藤润二、City Pop... 你的选择远不止于此。 应群众要求，整理了各种顶级动漫风格关键词，直接打开新世界大门。 来，给兄弟们盘一下~👇🏻

#前女友来了 真人感无敌的一张，仔细看细节 咒语： A natural grab shot of an alluring day-style woman hiking in the mountains, ponytail with bangs, flushed cheeks from exertion, in a sports bra top and leggings, pausing to look at the view, misty background with slight fog, captured mid-step for dynamic feel.

#前女友来了 咒语： A spontaneous photo of a sexy day-vibe woman on a subway train, long dark hair with straight bangs, fair complexion, wearing a cropped camisole and mini skirt, gazing out the window dreamily, motion blur from the moving train, reflective glass adding depth.

#前女友系列 咒语： A candid snapshot of a young woman with straight black hair, blunt bangs, large expressive eyes, and fair skin, standing casually in a bustling Tokyo street, wearing a low-cut crop top and high-waisted shorts, slight motion blur on the edges, natural daylight filtering through buildings, captured as if from a hidden angle with a smartphone camera.

前面给大家分享了很多图，很多推友问我怎么做的？ 其实AI 生图的底层框架只有三步： 主体 + 环境 + 风格。 其中“风格”，一语定乾坤。 我的风格库里整理了上百种风格关键词， 帮你少堆形容词、直接出好图，建议收藏。👇

很多朋友留言说，怎么同一个提示词，效果差那么多？ 如果你是用Nano Banana, 想要精准控图，就必须要用JSON prompt。（MJ相反，提示词过于复杂并不好） 关于人物“美不美”，这是一个很主观的标准，如果你通过"beautiful","stunning"等形容词来描述，其实不一定有效果 所以尝试给模型一些客观参考，比如让她的五官参考迪丽热巴、杨幂等等明星，效果会要好一些 P1为没有样貌参考，其它为有样貌参考 就用这张真实感拉满的图为例，外貌可以自己微调，咒语如下（包出的）： { "scene": { "description": "Young, attractive woman pausing mid-hike in a misty mountain forest, sweat beading on her skin.", "environment": "Cloud forest with light fog; soft greenery and mossy rocks disappearing into haze.", "mood": "Fresh, natural, energetic and charming." }, "output": { "aspect_ratio": "2:3 vertical EXACT", "target_resolution": "3072x4608", "safe_margins": "6–8% inner padding" }, "aesthetic": { "style": "Naturalistic outdoor portrait (DSLR look).", "look": "Cool-neutral palette; gentle filmic contrast; realistic wet-skin highlights; fog-softened background." }, "lighting": { "description": "Overcast sky as giant softbox; key from high-front camera-left with wrap; very light fill from camera-right; micro-speculars on cheekbones, nose bridge and collarbone.", "exposure": "Preserve highlight detail on moist skin; background ~0.7 stop darker for separation." }, "subject": { "demographics": { "age": "18–25", "gender": "female", "ethnicity": "East Asian", "beauty": "Youthful, healthy glow; balanced symmetry; refined features." }, "appearance": { "hair": "High ponytail; light, piecey bangs over the forehead that are ONLY slightly damp; a few wispy strands touch the cheek; slight humidity frizz.", "skin": "Natural pores with peach-fuzz; visible sweat beads on face/shoulders/clavicle; subtle rosy cheeks.", "makeup": "Minimal—hydrated base, soft lip tint, groomed brows.", "likeness_guidance": "Facial character INSPIRED BY Dilraba Dilmurat (迪丽热巴) and Yang Mi (杨幂)—large bright almond eyes with long lashes, defined straight brows, refined high nasal bridge with gentle tip, heart/oval face with soft jawline, fuller natural lips—DO NOT exactly replicate any specific person." }, "wardrobe": { "top": "Graphite-grey sports bra with knit micro-texture and wide straps." }, "accessories": { "gear": "Subtle backpack strap or translucent rain shell edge along subject's left side.", "jewelry": "None." }, "pose": { "type": "Chest-up close portrait (bust).", "framing": "Crop just above navel; both shoulders in frame; head near upper third; negative space to the left for fog.", "body_orientation": "Torso turned 15–20° toward camera-left; right shoulder slightly closer and lower by ~2–3 cm.", "head_orientation": "Head turned 20° toward camera-left; tilt (roll) −3°.", "gaze": "Eyes looking off-camera toward upper-left; relaxed expression.", "mouth": "Lips relaxed, slightly parted.", "hands": "Not visible." } }, "camera_lock": { "camera_height": "Upper-chest (sternum), ~10–12 cm below chin.", "camera_distance": "0.8–1.0 m from subject.", "camera_angle": "Yaw 0°, pitch −2°, roll 0°.", "lens": "50mm equivalent (acceptable 45–60mm).", "depth_of_field": "f/2.8 look—eyes/lashes/hair strands sharp; background softly blurred." }, "props_in_scene": { "background": "Soft bokeh foliage and fog; no signage or text.", "materials": "Moist skin, gently damp bangs, woven athletic fabric, fine fog particles." }, "camera_technical": { "requirements": "True facial proportions; crisp eyelashes and individual hair strands; readable sweat beads; no text/watermarks or AI glyphs.", "capture": "ISO 200–640; 1/250 s; WB 5300–5600 K; EXACT 2:3 vertical." }, "retouching": { "notes": "Preserve pores, droplets and wispy bangs; remove only transient blemishes; avoid plastic smoothing.", "grade": "Subtle cool-neutral grade; gentle S-curve; minimal vignette." }, "avoid": [ "Over-wet bangs (must be only slightly damp)", "Harsh flash or clipped highlights", "Aged/tired look—must remain youthful and pretty", "Pose/angle drift from locks", "Text/logos/watermarks", "Over-smoothing that erases texture" ] }

继续复刻 第三弹~ 这次挑得这张同样非常经典~ 1.从构图角度来说： 两个大叔在两侧构成视觉权重的对称平衡，让中间妹子显得更突出、更纯净 她的明亮肤色、制服颜色在中间形成视觉中心，对比两侧大叔黝黑的肤色，更进一步强化了主体焦点 2.从叙事角度来说： 制服象征纪律，环境却破旧凌乱，三人并坐，却无明显互动 妹子气质温柔，却身处权威符号之中 看似随意，实则矛盾共存，极具叙事张力~ 然后复刻难度也很大，为什么？ 因为原图妹子实在太好看了，很难超越😆尽力了兄弟们 老规矩，prompt在评论区，记得回来交作业~

复刻第四弹～ 你被逮捕了！ 这张比较难调的是腿部细节，很容易不自然。 MJ在生成人物时并不能真正理解人体结构或重力平衡，它只是根据大量图像的像素分布去猜测形态。 当涉及到腿部这种需要精确骨骼支撑、重心转移和透视变化的部位时，模型容易混淆姿态逻辑， 导致腿的比例、弯曲方向或接地方式出现不自然的结果。 老规矩，评论区完整prompt，记得回来交作业！ 友情提示，这张不要轻易尝试grok生成视频🙅🏻‍♂️

你的AI生图总是平平无奇？ 除了前面提过的艺术风格类型，还有一点你需要了解： 拍摄视角。 - GoPro view → 极限运动感 - Drone view → 史诗航拍感 - shot on 70mm → 好莱坞质感 - Bullet time → 黑客帝国特效 ... 下面的拍摄视角关键词， 让你的AI秒变专业摄影师。 收藏这条，拯救你的构图 🎬

完美复刻karin，可以起号了！😎 试了很多个模型seeddream，banana，recraft，只有midjourney才能找到这种日系朦胧感的精髓。 提示词如下，大家可以去试试： a dreamy intimate portrait of a beautiful realistic asian woman lying under a blanket, soft natural light, cozy warm bedroom, cinematic film grain, 35mm film aesthetic, shallow depth of field, soft focus, delicate skin texture, natural makeup, slightly messy hair, Fujifilm Pro 400H look, pastel warm tones, realistic lighting, subtle haze, gentle atmosphere, by Petra Collins and Ren Hang --ar 3:4 --v7 --style raw

继续复刻！ 这次挑选了一张karin点赞量最高的作品，复刻难度也不小，炼了一晚上😅。 主要难点在于整体氛围感的拿捏和女生的妆容，拆解下提示词重点（完整prompt放评论区，记得回来交作业）： 1.女生的妆造描述 一开始腮红老是打太重，跟年画娃娃似的，后面加了一句好多了：faint pink flush on the cheeks, soft and natural lighting 2.物理环境描述 居酒屋，被中年大叔环绕，货架等描述用以复现画面中的人群结构和社会氛围。 3.视觉风格描述 主要强调低对比度，复古氛围，35mm胶片的细腻颗粒感，给了一些风格参考森山大道和王家卫。 最后精选了4张，你更爱哪张？

这俩妹子身材巨好，跳得也很好看 唯一不足的就是： 每个舞蹈视频选的bgm很不搭 当舞蹈动作和音乐节奏高度同步时，我们的大脑会产生一种强烈的预期满足和时间对齐的快感 相反，如果不同步，大脑的预期被打破，就会带来不适感 视频来源：视频号-C颖Emliy

换一种摄影风格，OK吗？ 咒语： A captivating black and white close-up portrait of a beautiful young Asian woman with flawless skin, short bob haircut featuring straight blunt bangs framing her forehead, eyes gently closed with long lashes, subtle pouty lips, serene and introspective expression, wearing a form-fitting deep V-neck black ribbed sweater that accentuates her, a delicate thin necklace with a small pendant, dramatic side lighting from the left creating strong shadows and highlights on her face, neck, and shoulders, high contrast chiaroscuro style, minimalist studio background with soft gradient from light to dark, photorealistic, vintage film grain effect, high detail, emotional and mysterious atmosphere.

黑白风格的大家很喜欢啊，再来个吧 记得交作业，24小时后选出高质量作业公示🫶 咒语： a woman standing confidently in a minimalist studio, shot from a slightly low angle, half-body portrait, her head tilted slightly down with a calm and enigmatic expression, middle-parted braided pigtails with a few loose strands falling over her face, subtle wet hair texture, wearing a tight black sleeveless dress and black leather gloves, holding a leather belt loosely in front of her body — dramatic single key light from upper right, high contrast, black and white tones, strong rim light accentuating her curves, plain textured wall background, cinematic noir atmosphere, fine art fashion photography, inspired by Helmut Newton and Peter Coulson, shallow depth of field, smooth gradients, elegant and powerful mood

#前女友系列 咒语： A candid snapshot of a young woman with straight black hair, blunt bangs, large expressive eyes, and fair skin, standing casually in a bustling Tokyo street, wearing a low-cut crop top and high-waisted shorts, slight motion blur on the edges, natural daylight filtering through buildings, captured as if from a hidden angle with a smartphone camera.

前面给大家分享了很多图，很多推友问我怎么做的？ 其实AI 生图的底层框架只有三步： 主体 + 环境 + 风格。 其中“风格”，一语定乾坤。 我的风格库里整理了上百种风格关键词， 帮你少堆形容词、直接出好图，建议收藏。👇

#前女友来了 咒语： A spontaneous photo of a sexy day-vibe woman on a subway train, long dark hair with straight bangs, fair complexion, wearing a cropped camisole and mini skirt, gazing out the window dreamily, motion blur from the moving train, reflective glass adding depth.

Chroma Fade 记得评论区交作业，24小时后放榜优秀作业😊 咒语： young beautiful woman, calm confident expression, subtle smirk, short white bob haircut, black lipstick, sharp eyeliner, tattoos on neck and arms, wearing sleeveless black turtleneck top, gothic elegant fashion, multiple ear piercings background dark studio, cinematic rim lighting, high contrast light and shadow, smoky atmosphere, black and white photography, professional portrait, half-body shot, soft diffused lighting, moody tone, ultra-detailed, realistic, 8k, masterpiece

#前女友来了 咒语： A stealthy candid capture of a Asian beauty at a night market, straight hair with bangs, rosy cheeks, wearing a deep-neck tank top and jeans, browsing stalls with a curious gaze, low-light with neon signs blurring in the background, grainy texture like from a phone in dim conditions.

Unspoken Whisper 咒语： A young beautiful woman in dramatic chiaroscuro lighting, softly tilting her head back, half-closed eyes, melancholic expression, cigarette between her lips with thin smoke curling upward, shoulder bare, face partially in shadow, hair slightly messy and covering part of her eye, illuminated by a single narrow beam of light from the side smoky atmosphere, minimal background, cinematic noir mood, strong contrast between light and darkness, high grain texture, black and white photography, shot with a vintage lens, shallow depth of field, soft focus, film grain aesthetic, evocative emotional tone, timeless elegance style of 1960s film portrait, Rembrandt lighting, artistic sensual glamour

#赛博非诚勿扰 🌸【1号女嘉宾 · Rachel】 上海姑娘，牛筋大学毕业。她喜欢咖啡拉花、花店香气，还有听人讲故事。 她说自己是ENFJ——温柔但不软弱，懂得照顾人，也希望被认真对待。 她的朋友说，Rachel总能在混乱的世界里，给人一点安静的温度。 💬Rachel想问你两件小事： 1.如果我说“没事”，你真的会信吗 2.你最近在干嘛 💗爆灯也行，撩一句也行。反正她在等。 👇她录了一段VCR，说给懂的人听。 心动密码： a young beauty creamy-tone minimal look, beige knit sweater and white jeans, soft straight hair, dewy base and clear gloss, airy living room daylight, 50mm

你们可能不知道出片前我要经历些什么..