LowKey

宁可一思进，莫在一思停。

[林悦己Cheer @cheerselflin]

生成一段 [总时长] 秒、[画幅] 的 [地点／主题] 旅行 Vlog，共 [镜头数] 个镜头，全程只出现同一位 [人物]

人物锁定：[脸、发型、体型、肤色、妆容] 全程一致
固定造型：[服装、鞋、包、配饰]
条件道具：[道具] 只在 [指定场景] 出现

统一质感：[光线、色彩、手机／相机质感、真实感]
剪辑节奏：快速切换，每个镜头只完成一个主动作和一个自然反应

[镜头 01｜开始时间–结束时间｜镜头目的]
场景：[地点与必要环境]
动作：[一个主动作 → 一个自然反应]
摄影机：[景别、机位、单一运镜]
结束状态：[人物位置、朝向、手中道具和情绪]

[按相同格式继续后续镜头]

声音：[每个场景对应的环境声、动作声和音乐]
禁止：[变脸、换装、额外人物、道具无故消失、过度磨皮、棚拍感、CGI 感]

最终画面：[最后一帧的人物动作、表情、道具和背景]

[Ai船长 @Liam116888]

强烈建议兄弟们都去看这部Netflix高分纪录片：《聪明生活经济学》

看完彻底刷新我的金钱认知，穷人跟富人的最大差距，就是金钱思维！

💰不要死存钱，而是精简开支、拒绝即时享乐、理性消费、规划资产。

静下心来问问自己，未来到底想过什么样的人生？

想靠近财务自由，先从重塑金钱观开始。

#聪明生活经济学 #金钱观 #理财思维 #财务自由

[Faptuary @faptuary]

Scarlett Johansson being one-and-done with nudity but producing some of the most gratuitous nude scenes through that is one of the most bittersweet things there is.

[Zhen Zhu @ZhenZhu200]

【AI数字人运动教练 02：角色一致性】
AI数字人运动教练生产运动内容视频最大的问题是，角色漂移。

我没有停在一张“漂亮正脸”，而是给网球教练 Da‑Yul 做了一套 Character Bible：转身多视角、头部多视角、服装细节、动作演示。

这些资产图帮助你去确认这个教练的每一个细节，让她更加生动立体，以及看到的姿势、肌肉动作、四肢和五官比例是否合适，锁死教练的所有特征细节，保持一致性。

How：做一张你的角色图全身照，然后把我下面的四张图丢给agents，说：“基于我的角色图，只参考下面四张九宫格图的布局和方位，给我全面定义角色的转身多视角、头部多视角、服装细节、动作演示九宫格。”

[Kōda @aimikoda]

GPT Image 2 Character Sheet Prompt

A lot of people have been asking for this character sheet prompt, so I'm sharing it here. I've actually shared it under a few of my previous posts before but it seems like most people missed it.

It works from a reference image you provide and I originally designed it specifically to preserve the visual style of referenced Midjourney images as closely as possible.

You can change the aspect ratio however you like. Also it would work with Seedream 5 and other image models.

Prompt:
Create a 2:1 premium cinematic CHARACTER IDENTITY BOARD using Image A as the sole character and style reference.

Match Image A exactly: same identity, proportions, hairstyle, outfit, colors, visual medium, linework, shading and texture. Do not redesign or reinterpret. Correct obvious anatomy or generation errors like hands without changing the intended design. without sword and yellow dots.

REFERENCE LOOK LOCK
First analyze Image A and establish a strict visual identity lock from all observable character traits, proportions, design details and stylistic rules. Apply this locked identity consistently across every angle, pose and expression. Nothing may drift, relocate, disappear, duplicate or be invented; only perspective-dependent visibility may change.

COMPOSITION
Use a clean minimal background color chosen to contrast clearly and elegantly with the character’s palette. Avoid grids, blueprints and standard turnaround-sheet design. Never overlap or crop the studies.

RIGHT 25%: HERO PORTRAIT
The rightmost 25% must be a full-height, edge-to-edge close-up portrait completely filling that area with no margins or empty space. Match Image A’s identity, art style, lighting, colors and background treatment exactly. Keep the face and defining head silhouette clearly readable. Place no text or other studies over it.

LEFT 10%: EXPRESSIONS
Reserve the leftmost 10% for exactly four expression portraits. Place the four clearly separated expression portraits vertically.

Give each portrait a strong, distinctly different emotion and head direction: front-facing, left three-quarter, right three-quarter and side profile. The gaze should naturally follow each head angle. Do not repeat the same angle or expression. Preserve the exact same facial identity, proportions, features and art style in all four portraits.

CENTER 65%: BODY STUDIES
Use the central 65% for:

- one large off-center hero full-body pose, facing directly toward the viewer with the head straight, eyes clearly visible and the face sharp, unobstructed and easy to recognize
- back full-body view in a different pose
- side full-body view in a different pose
- one top-down body angle
- one low-angle body angle

Keep all views fully visible, clearly separated and consistent with the locked reference identity.

No text, arrows, labels, handwritten notes, annotation marks, silhouettes, detail studies, seated poses, leaning poses, crouching poses, environments, extra characters, alternate costumes, logos, watermarks, style drift, malformed anatomy, inconsistent features, duplicated details, obscured faces or overlapping figures.

[LOOPY @VibeEverything]

사용한 모델: Nano Banana Pro

[프롬프트]
Create a photorealistic three-panel character reference sheet based strictly on the uploaded reference image. The uploaded image is the single source of truth for identity, age, facial structure, hairstyle, body proportions, skin tone, clothing design, fabric, colors, and accessories. Do not redesign or reinterpret.

LAYOUT — Horizontal landscape composition, read left to right, arranged as three vertical panel areas of equal height and equal width. All three panel areas must have exactly the same width, each occupying one third of the total image width. The first two panels are full-body views, and the third is a face / upper-body close-up. The three panel areas must form one continuous, seamless composition with no visible divider lines, borders, frames, gutters, gaps, spacing, or changes in the background between panels.

PANEL 1 — FULL-BODY FRONT VIEW (HEAD OMITTED)
Standing straight in a neutral pose, arms naturally at the sides, facing directly toward the camera. The outfit is fully visible from the neckline down to the footwear. First establish the figure at exactly the same full-body scale and framing as in Panel 2, then omit only the head area above the neckline, leaving normal empty space where the head would have been. No decapitation effect, no wound, no severed neck, no surrealism, and no anatomical damage — a clean character-design presentation in which the head information has been intentionally removed. Never enlarge, zoom, reframe, shift, or re-crop the body to compensate for the missing head.

PANEL 2 — FULL-BODY REAR VIEW
The exact same character seen from directly behind, in the same neutral pose, arms at the sides. Clearly show the back silhouette, the hairstyle from behind, the garment construction and seams, the lower-body garment, and the footwear, exactly as in the reference.

PANELS 1 & 2 — SCALE-LOCKED
Shot consecutively as if from a tripod that never moved: identical focal length, camera height, camera distance, subject scale, and lighting. Identical on-screen shoulder width, torso length, leg length, shoe size, shoulders-to-floor distance, and side margins. Both figures share one identical horizontal ground baseline and one identical horizontal shoulder line, and each figure occupies the same percentage of the panel height. No zoom, crop, vertical shift, or difference in body size between them. Near-orthographic, technical feel with minimal perspective distortion.

PANEL 3 — FACE / UPPER-BODY CLOSE-UP
Head-and-shoulders portrait of the exact same character, cropped from the upper chest upward. Turn the head slightly to the subject’s right, creating a subtle three-quarter angle while keeping the face mostly toward the camera. The subject’s left cheek and left side of the face must be visibly more exposed than the right side, with the nose pointing slightly toward the subject’s right rather than straight at the camera. Keep both eyes clearly visible and maintain direct eye contact with the camera. Do not use a perfectly frontal, symmetrical face angle. Maintain a serious neutral expression. Keep the entire head, hairstyle, chin, and hair outline fully inside the frame with a small amount of space above the hair. Preserve the exact identity, age, facial structure, hairstyle, skin texture, wrinkles, and natural asymmetry of the  reference.

CONSISTENCY
Identical character identity, age, skin tone, hairstyle, body proportions, wardrobe (every garment, its cut, and all accessories exactly as in the reference), fabrics, colors, lighting, and color grading across all three panels. This is a technical visual reference / costume documentation sheet, not a creative reinterpretation.

BACKGROUND
Neutral mid-grey (#808080) seamless background extending continuously across the entire image without any interruption or panel separation. Identical background color and brightness across all three panel areas.

LIGHTING & RENDER
Clean soft even lighting, large diffused key light with gentle fill, soft natural shadows, no harsh highlights, true-to-life skin tones, neutral white balance, minimalist high-fashion editorial look, polished modern professional model-sheet aesthetic, shot on a full-frame camera with an 85mm lens, deep controlled depth of field with the entire figure in focus, crisp tack-sharp detail, high dynamic range.

NEGATIVE
No text, labels, typography, logos, measurements, props, furniture, environmental objects, decorative graphic design, divider lines, borders, frames, gutters, panel gaps, or visible seams between panels. No unequal panel widths, and no panel wider or narrower than the others. No rim light, colored lighting, harsh spotlight, or theatrical lighting.

[LudovicCreator @LudovicCreator]

🎨 CINEMATIC WARTIME CHRONICLE 🎨

Prompt :

A Cinematic Wartime Chronicle featuring [SUBJECT], portrayed through grainy, scratched film stock typical of wartime newsreels. Utilize faded military-inspired hues of [COLOR1] and vintage-muted [COLOR2] tones, complemented by gentle camera shake and reel flicker

Check ALTS

[John @johnAGI168]

像变魔术一样 Amazing🌟🔥

抖音上刷到一个很酷的视频，使用 seedance 2.0 复刻了一个 。

prompt 👇

【风格】东方奇幻控火武术短片（Oriental Fantasy Live-Action），真实摄影（Photorealistic），电影级安全火焰特效（Cinematic Fire VFX），高对比夜景，白热火芯、橙黄外焰、红褐烟雾、真实火星与动态火光，IMAX构图，4K高清，全片实时速度、无慢动作
【时长】15秒
【场景】开场是青绿色厚窗帘前的简洁现代室内；随后通过动作匹配硬切，进入夜间中式木构建筑前的宽阔石板院落。院落背景低调昏暗，中央木门、台阶、廊柱和石栏位置全片固定，周围无人、无文字
【角色】角色A 📷图片1 ，人物身份和脸部严格遵循参考图。开场穿简洁黑色短袖；切换后穿黑色多层古风武者服装，宽袖、宽下摆、腰部束带和高束长发。古风造型一经出现，全片保持不变

【核心道具】一根固定长度的深色细长表演杆，只有一端带电影级安全火焰特效；角色A始终双手牢固握住杆身，杆身不能消失、缩短、弯曲或脱手

[00:00-00:01.20] 镜头1：横臂瞬变（Match Cut Transformation）
【机位】固定正面中近景（Locked Medium Close-Up），角色A站在青绿色褶皱窗帘前，右后方硬光在窗帘上投出清晰人影。
【动作】角色A直视镜头露出浅笑，随即低头看向右手；右拳从腰侧迅速绕到左肩，再以很快速度横扫胸前。

【爆点】右前臂扫到画面中线、运动模糊最强的一帧，执行动作匹配硬切（Action Match Cut）：窗帘、现代服装和室内灯光一帧消失，瞬间切成夜间古风石板院落和黑色古风武者造型；右手边同时爆开白热火焰、橙色火舌、火星和灰烟。
【固定】不是衣服连续生长，不是身体变形；人物脸部、手臂位置和横扫方向在切换前后准确衔接。
【音效】手臂破风声接低频重拍，变装切点同步一声火焰轰鸣。
[00:01.20-00:03.00] 镜头2：伸掌控火（Fire Control Close-Up）
【机位】保持固定中近景，人物比例和前一镜头一致，焦点锁定角色A的右手与掌前火团。
【动作】角色A延续横扫动作，把右臂完全伸向画面左侧，手掌突然张开，五指清楚分离；角色A低头看向掌前火团，嘴唇闭合，眉眼收紧。

【特效】一团白热核心、橙黄外焰的火焰停留在手掌前方左侧，连续收缩再膨胀；红褐烟雾向上翻卷，细小火星向下散落。火光从左下方动态照亮角色A的脸侧、手掌和黑衣边缘。
【固定】火团来自电影视觉特效，不接触皮肤；不出现第二人物，不生成具体火焰生物。
【音效】火焰持续呼啸、细碎火星爆裂声，配乐节拍逐渐加快。
[00:03.00-00:04.20] 镜头3：横扫成环（Horizontal Fire Ring）
【镜头切换】火团仍位于画面左侧时，硬切到正面固定大全景（Wide Locked Shot）。
【画面】角色A站在古建庭院中央，双腿大幅分开并降低重心，双手握住单端燃烧的深色长杆。

【动作】角色A以腰胯带动长杆，从左后方向右侧完成一次快速水平旋转；燃烧端先划出一条橙黄色月牙，随后扩张成围绕角色A下半身的巨大水平火环。
【特效】火环具有白热核心和橙红边缘，火焰经过后仍短暂悬留；环外飞散火星，灰黑烟雾滞后上升并逐渐遮住角色A下半身。
【收束】火环快速碎裂成地面余火，浓烟填满画面中下部。
【音效】长杆破风声沿左右声道移动，火环闭合时落下一记低频鼓点。
[00:04.20-00:08.40] 镜头4：竖环升空（Vertical Fire Ring，Continuous Action）
【转场】利用浓烟遮住角色A，在烟幕最厚时硬切重置动作；场景和人物身份不变。
【00:04.20-00:05.40】角色A重新直立，长杆斜指左上；双手把燃烧端从左上带到左下，再贴近石板地面从左向右横扫，随后由右侧高举过头顶。燃烧端在竖直平面画出一个完整圆形火环，角色A始终位于圆心。

【00:05.40-00:06.70】火环坍散成地面火墙和浓烟。角色A沉入低蹲，把燃烧端从腰后绕到胸前右侧，再从低位起身，将长杆向头顶抡起；宽袖、长下摆和高束长发顺着旋转方向甩开。
【00:06.70-00:08.00】角色A宽步旋身，完成一次短促、轻盈的腾空动作；燃烧端沿人物外侧高速画圆，月牙火弧迅速膨胀成包围全身的巨大竖直火环。
【00:08.00-00:08.40】角色A自然下降，双脚接触石板后屈膝缓冲；巨大火环同步断裂成火星、地面余火和厚重烟雾。
【机位】全段保持正面固定中远景，不推拉、不绕拍。
【固定】腾空时间短促，必须看见起跳、下降、脚掌落地和屈膝缓冲；人物不能在空中静止，长杆不能与双手分离。
[00:08.40-00:10.70] 镜头5：俯瞰画圆（Top Shot Fire Circle）
【镜头切换】人物落地、烟雾铺开时，硬切到90度垂直俯拍（Top Shot）。
【动作】角色A位于圆心附近，保持低身姿态旋转长杆；燃烧端依次经过人物右侧、画面下方和人物左侧，在与地面平行的水平面画出明亮U形火轨，再接近完整圆形火圈。

【特效】最新火头呈白黄色，旧火迹逐渐由橙红变成暗红余烬，随后熄成灰黑烟圈；主火头始终固定在长杆同一端。
【运镜】摄影机持续垂直升高并缓慢拉远，角色A、火圈和烟雾逐渐缩小，四周石板沉入黑暗。
【固定】火圈与地面平行，不倾斜、不立起；不出现火龙、凤凰、火鸟或悬浮光球。
【音效】火圈经过画面四个方向时配连续环绕破风声，拉远时配乐短暂降低。
[00:10.70-00:15.00] 镜头6：斜劈起火墙（Low-Angle Fire Wall Finale）
【镜头切换】俯拍画面只剩一个小火点时，硬切到低机位广角荷兰角（Low-Angle Dutch Shot）。
【画面】角色A站在夜间石板院落中央偏左，双手把同一根燃烧长杆高举过头；背景左后方是暗色木构门窗，右后方只有少量暖色灯点。

【00:10.70-00:12.00】角色A向镜头方向压进半步，转动腰胯并降低重心，把燃烧端从头顶沿身体左侧斜向挥到镜头左下方；火焰逼近镜头，形成短暂黄白过曝和大片橙色失焦散景。
【00:12.00-00:13.30】角色A抬臂完成收势；燃烧端的挥动轨迹在画面上方扩张成横跨左右的巨大半圆火弧，画面下方中央和右侧同时升起粗大的翻卷火墙。
【00:13.30-00:14.50】火墙内部呈黄白色，外缘呈橙红色，表面不断翻出细碎火舌；黑烟向火弧下方堆积，火星向上散落。角色A稳定站在半圆火弧内部，衣摆受热流和动作惯性轻微摆动。
【00:14.50-00:15.00】火墙快速回落，白热核心缩小，橙红余火和黑烟占据画面；角色A保持收势姿态，火弧在头顶逐渐断裂，最后一簇火星落下完成收尾。
【固定】从斜向挥杆到火墙展开保持同一连续镜头，火焰遮挡不能触发换景；随机火舌不能变成具象动物。
【音效】斜向挥杆配沉重破风声，火墙升起配低频火焰轰鸣；最后火势回落时配乐停止，只剩火星落地声和烟火余响。
【音响】
全片无对白、无旁白、无人物喊叫。使用低沉东方打击乐、持续环境低频、长杆破风声、火焰呼啸声、火星爆裂声和烟火闷响；硬切、火环闭合、腾空和最终火墙分别与音乐重拍同步。音效必须跟随燃烧端的真实运动方向，不提前出现。
【角色与道具固定】

角色A的脸部、身份和肤色严格遵循@图片1；现代服装只存在于开场，1.20秒切换成古风造型后不得恢复或改变。
深色长杆全片长度、粗细和颜色不变；角色A始终双手握住杆身，只有同一端燃烧。火焰必须由燃烧端的运动轨迹生成，不得从角色A身体、头发、手掌或长杆另一端冒出。

【全片禁止】
连续融化式变装、衣服从皮肤生长、背景渐变、无剪辑换景、人物脸部变化、身份漂移、服装颜色变化、发型变化、第二人物、群众、动物、火龙、火凤凰、火焰怪物、悬浮光球、字幕、片名、台标、logo、水印、乱码。
禁止长杆变短、变粗、弯曲、变成软绳、复制成两根、穿过身体、脱离双手、握点漂移、双端燃烧、火焰离开杆端独立飞行。禁止手指融合、手脚增减、关节反折、身体重影、人物瞬移、角色出画后重新出现。

禁止长时间悬空、空中静止、缺少落地过程、脚掌穿过地面、落地后身体无缓冲。禁止火环平面突然翻转、火圈与长杆脱离、火焰一帧消失、烟雾瞬间重置、卡通火焰、动漫线条、塑料皮肤、大面积纯白过曝。
禁止镜头抖动、甩镜、冲击变焦、无故旋转、机位漂移、错误运镜、反射镜头和无关转场。

[Arvin @Arvin010717]

必须严格按照每一格中的：人物姿态、攻击方向、受力位置、箭头轨迹、动作顺序、镜头距离进行生成。 分镜图中的： 箭头、 编号、 轨迹线、 分析标记、 导演辅助元素仅作为内部动作参考。 最终视频画面中禁止出现：箭头、 文字、 编号、 速度线、 UI、 分析图层、 运动轨迹残留。 输出必须是： 纯净电影画面。 白衣剑客：白色水墨、飞白残影、高速剑术、长距离切线位移。 青衣拳客：深青浓墨、徒手压迫、肘击、肩撞、近身抢位。 禁止修改角色颜色，禁止角色互换，禁止新增武器。 视频节奏要求： 每1秒至少出现2-3次有效攻防转换， 每一次攻击都必须被拆解成：发力、接触、受力、回弹、再进攻。 动作不能一闪而过，必须看清动作逻辑，但整体速度仍然极快。 严格执行分镜中的： 箭头轨迹、进攻路线、位移方向、跳跃方向、身体旋转方向。 重点表现：贴身短打、剑路变线、翻身反刺、贴地滑步、连续抢位、身体碰撞。 镜头必须严格贴合分镜：近景跟拍、中近景压迫感、贴身环绕、低机位冲杀、高速方向切换。 禁止：角色漂移、自动补帧、镜头乱飞、动作重复、AI自动编动作、慢动作、特效化仙侠。 水墨表现必须动态化：白衣角色形成飞白墨痕，青衣角色形成浓墨爆散，碰撞点出现泼墨炸裂，高速移动产生书法式拖墨。 音频要求：无背景音乐，无配乐，无电子音效，无史诗BGM。 只保留自然环境声与战斗声：悬崖狂风、衣物破风、脚步摩擦、身体碰撞、拳击闷响、剑锋破风、金属碰撞、碎石断裂、呼吸声、近距离受力震动声。   电影级水墨动画，真实宣纸纹理，狂风云海，悬崖碎石，极强东方武术暴力美学。

[Saul Goodman @Goodmanprotocol]

Mid-century travel picture-book illustrations of American cities

GPT Image 2

Prompt:
[DESTINATION] = [CITY NAME]

Create a charming vertical 4:5 mid-century travel picture-book illustration of [DESTINATION], inspired by playful 1950s–60s European city books.

Depict the city as an affectionate, imperfect travel memory—not a map, postcard, architectural rendering, landmark collage, or cinematic panorama.

Include:
• 1 distinctive primary landmark
• 1 smaller secondary landmark
• 1 distant/partially hidden local feature
• 1 recognizable local vehicle
• 3–5 subtle everyday city details
• 5–6 ordinary people doing simple local activities

Compress these naturally into one imaginary street, plaza, park, or waterfront. Keep the primary landmark slightly off-center, with open foreground, shallow theatrical depth, gentle curves, and generous museum-white negative space. The city should be recognizable without feeling geographically accurate.

STYLE:
Hand-drawn picture-book illustration with loose charcoal/navy ink, trembling contours, crooked architecture, uneven proportions, sparse windows, simplified silhouettes, occasional double strokes, and minimal detail. Use the fewest lines necessary. Slightly awkward, intelligent, warm, humorous, and handmade.

PEOPLE:
Simple expressive figures with oversized heads, thin legs, narrow bodies, tiny faces, tilted accessories, and awkward but readable gestures. Show ordinary local life, not celebrities, models, stereotypes, or tourist clichés.

COLOR:
Use a restrained 6–7 color destination-specific palette with flat matte gouache shapes, uneven brush edges, white gaps, slight misregistration, and occasional color spilling beyond ink. Mostly unpainted sky/paper. No gradients or realistic shading.

MEMORY DISTORTION:
Gently compress geography and distort scale: distant landmarks may appear nearby, roads may curve strangely, buildings may overlap, and people/vehicles may be slightly oversized or undersized. Keep it balanced, affectionate, and culturally respectful.

OPTIONAL TEXT:
“[DESTINATION]” in small uneven hand-drawn capitals near the bottom. No slogans, dates, labels, logos, or promotional text.

AVOID:
Photorealism, map-like accuracy, landmark checklists, aerial views, cinematic lighting, technical perspective, dense crowds, detailed ornament, polished vector art, CGI, 3D rendering, gradients, sepia/beige paper, artificial aging, distressed textures, and hyper-detailed watercolor.

Overall feeling: [DESTINATION] remembered lovingly-recognizable, spacious, human, humorous, slightly improbable, and beautifully imperfect.

[Saul Goodman @Goodmanprotocol]

Seedance 2.5 made this feel like a forgotten European documentary from the 1980s. You’d swear it was real. 🎞️✨

Prompt:
Create a 30-second ultra-photorealistic early-1980s French countryside documentary, shot on vintage Kodak 16mm. It must look like genuinely forgotten archival footage discovered decades later—not modern footage with a vintage filter.

STYLE:
Faded warm Kodak colors, organic 16mm grain, subtle light leaks, soft highlight bloom, slight gate weave, exposure fluctuations, natural motion blur, imperfect focus, realistic skin texture and period-authentic optical imperfections. Warm golden afternoon sunlight. No digital sharpness, AI smoothness, glossy cinematic grading or commercial polish.

CAMERA:
Shoulder-mounted handheld 16mm documentary camera operated by a real human. Subtle shake, imperfect framing, focus hunting, natural camera lag, occasional reframing and slight exposure breathing. Observational and spontaneous. Nobody acknowledges the camera.

0–5s:
A slim, naturally beautiful blonde European woman in her early 20s walks through a quiet cobblestone street in a small French village. She wears a simple beige linen midi dress and brown leather sandals, carrying a woven basket filled with flowers and bread. Her hair is loosely tied with flyaway strands. No makeup, no posing. She trails her fingers along old stone walls while glancing at flower-filled windows. Close handheld tracking shot.

5–10s:
The camera falls slightly behind as she enters a cozy village square. A baker carries fresh baguettes, an elderly couple chats on a bench, colorful bicycles lean against ivy-covered buildings, and a small golden retriever crosses the camera path. The camera instinctively shifts to avoid the dog, creating an authentic human reaction. She continues walking naturally.

10–15s:
Outside a tiny flower shop, a little girl accidentally drops a bouquet of wildflowers. The woman gently kneels, gathers the flowers and returns them with a soft spontaneous smile. The child smiles shyly and runs toward her mother. Completely unstaged; nobody looks at the camera.

15–20s:
She continues through rustic stone streets with wooden shutters, laundry, flower boxes and small market stalls. A breeze moves her dress and hair. She pauses briefly at a bakery window, looking at fresh bread. The handheld camera drifts and corrects naturally.

20–25s:
She leaves the village and walks toward a quiet overlook. The camera follows from behind as the scene opens into rolling green vineyards and distant hills beneath golden afternoon light. Birds cross the sky.

25–30s:
At an old iron railing, she rests both hands and watches birds over the valley. Wind moves loose hair across her face; she tucks it behind her ear and smiles subtly. She walks away. The camera pauses, lets her become smaller in frame, then gently follows.

CONTINUITY:
Same woman, blonde hair, beige linen dress, basket and natural appearance throughout. Authentic early-1980s environment with no modern objects, clothing, cars or architecture.

FINAL LOOK:
True archival 16mm documentary realism, Kodak film grain, faded colors, realistic skin pores, organic exposure, subtle flicker, light leaks, gate weave, imperfect focus and authentic handheld motion. No AI beauty smoothing, plastic skin, excessive depth of field, slow motion or modern cinematic polish.

ASPECT RATIO: 16:9

[hedoήist @hedo_ist]

Wow, What do you think?：

This may have been made with Seedance 2.5, but that's not the key point.

Watch the credits: the team is Chinese.

While we’re still comparing models, Chinese creators are already building the full AI filmmaking pipeline—faster, cheaper, and better than most of us realize🤯

[李岳 @liyue_ai]

运镜提示词不会写？
试试我这个运镜Skill吧，
一键直出所有图片的运镜，
请欣赏成片《东方未来纪元》。

Skill地址：https://t.co/ZNawygNoAO

[tomasz.golonko @TomaszGolonko]

Wow. Maciej Dobrodziej - twórca cyfrowy „ożywił” jedno z najsłynniejszych zdjęć Warszawy. Fotografie biegnącej w deszczu dziewczyny na ulicy Puławskiej zrobił Zbigniew Siemaszko w 1968 roku.

Ciekawostka: po latach udało się odnaleźć panią Grażynę, która jest bohaterką zdjęcia. Rozpoznała siebie widząc.. fotografię na Fb.

[ヤノ @Ryuki_Yano]

Higgsfildで公開されている95分の映画のプロジェクトを読んだ結果、最もキャラクターシートとして使い勝手がいいのはこれなんじゃないかという結論に至りましたので公開します。使いたい方はどうぞ。

以下、キャラクターシートのプロンプトになります👇️

{{CHARACTER REFERENCE}} を唯一のキャラクターデザイン参照として使用し、
顔、髪型、髪色、肌、年齢感、体格、身長比率、衣装、アクセサリー、装備、素材、色、ディテールを正確に維持した、
プロフェッショナルなCHARACTER IDENTITY SHEET / CHARACTER BIBLEを制作する。

これは新しいキャラクターデザインを作るための画像ではない。
参照キャラクターの外見情報を、制作現場で再利用しやすいように整理した資料画像である。

背景は完全にクリーンな白〜ごく薄いニュートラルグレー。
高級ファッションブランドのキャスティング資料、映画制作のキャラクターバイブル、AAAゲームのキャラクター資料を組み合わせたような、
洗練され、情報整理されたミニマルなレイアウト。

横長の1枚のキャンバス。
画面を大きく「左：BODY / INFORMATION」「右：FACE IDENTITY」の2領域に分ける。

━━━━━━━━━━━━━━━━━━
【LEFT AREA：CHARACTER INFORMATION + BODY】
━━━━━━━━━━━━━━━━━━

画面左上には十分な余白を設け、
キャラクター情報をシンプルなサンセリフ体のタイポグラフィで記載する。

NAME:
{{NAME}}

HEIGHT:
{{HEIGHT}}

VOICE:
{{VOICE}}

CHARACTER:
{{PERSONALITY}}

必要以上に文章を増やさず、
映画制作資料のような簡潔で美しい情報設計。
黒またはダークグレーの文字。
装飾的なUI、SF HUD、派手なグラフィックは禁止。

その下に、同一キャラクターの身体・衣装確認用ビューを2体配置する。

LEFT BODY 01:
正面、完全なフロントビュー。
ニュートラルな直立姿勢。
腕は身体からわずかに離す。
左右対称に近い自然な立ち姿。
衣装、体格、シルエット、脚の長さ、肩幅、腰位置、アクセサリーが確認できること。

LEFT BODY 02:
完全な背面ビュー。
正面ビューとまったく同じ身長、縮尺、姿勢、身体比率。
背面の衣装構造、髪で隠れない背中部分、装備、靴などが確認できること。

この2つのBODY VIEWでは顔を描かない。
頭部そのものを不自然に消去するのではなく、
首から上をレイアウト上で自然にフレームアウトさせ、
BODY / COSTUME REFERENCEとして成立させる。

正面と背面を同じ床位置、同じ縮尺で並べる。
ポーズを付けない。
モデル立ち、アクションポーズ、歩行ポーズは禁止。

━━━━━━━━━━━━━━━━━━
【RIGHT AREA：FACE IDENTITY】
━━━━━━━━━━━━━━━━━━

画面右側はキャラクターの顔の同一性確認専用エリア。

最も重要なのは、
大きく配置された正面のFACE CLOSE-UP。

RIGHT MAIN PORTRAIT:
キャラクターの顔を大きく画面いっぱいに見せた、
真正面に近い超高精細バストアップ / 顔のクローズアップ。

頭頂部から肩付近まで。
真正面。
カメラ目線。
ニュートラルで自然な表情。
過剰な笑顔や演技は禁止。

目、眉、鼻、唇、輪郭、頬骨、顎、耳、
髪の生え際、前髪、髪型、
メイク、傷、ピアス、眼鏡など、
キャラクターを識別するすべての特徴を明確に描写する。

この正面顔を、このキャラクターのPRIMARY FACE IDENTITYとして扱う。

さらに正面顔の周囲または右側に、
小さめの顔ビューを2つ配置する。

FACE VIEW 02:
3/4 VIEW。
顔を約30〜45度横に向けた斜め顔。
正面顔と完全に同じ人物。
自然なニュートラル表情。
鼻の立体感、頬骨、顎、耳、髪型の奥行きを確認できる。

FACE VIEW 03:
TRUE SIDE PROFILE。
完全な横顔。
正確な90度プロフィール。
正面顔と完全に同じ人物。
鼻、額、唇、顎、後頭部、首のラインを明確に確認できる。

3つの顔ビューすべてで、
顔の骨格、目の大きさ、目の間隔、眉、鼻、唇、
顎、耳、肌、髪型、髪色を完全に統一する。

別人化禁止。
ビューごとの美化・年齢変化・メイク変更禁止。

━━━━━━━━━━━━━━━━━━
【LAYOUT】
━━━━━━━━━━━━━━━━━━

左側：約45%
・上部：キャラクター名とプロフィール
・下部：顔をフレームアウトした正面BODY VIEW
・その隣：顔をフレームアウトした背面BODY VIEW

右側：約55%
・最も大きい正面FACE CLOSE-UPを主役として配置
・その周囲に3/4 FACE VIEW
・TRUE SIDE PROFILEを補助的に配置

正面FACE CLOSE-UPは、
他のどのビューよりも明確に大きくする。

すべてのビューの間に十分な余白を設ける。
画像同士を重ねない。
グリッド感を強く出しすぎず、
高級なエディトリアルレイアウトとしてまとめる。

━━━━━━━━━━━━━━━━━━
【CONSISTENCY】
━━━━━━━━━━━━━━━━━━

すべて同一人物。
同じ年齢。
同じ顔。
同じ髪型。
同じ髪色。
同じ肌。
同じ身体比率。
同じ衣装。
同じアクセサリー。
同じ素材。
同じデザイン。

デザイン変更禁止。
衣装変更禁止。
別衣装追加禁止。
新しいアクセサリー追加禁止。
顔の再デザイン禁止。
髪型変更禁止。
体型変更禁止。
色変更禁止。

キャラクターを5人描くのではなく、
「1人のキャラクターを5つの資料ビューで表示している」ことを明確にする。

━━━━━━━━━━━━━━━━━━
【VISUAL QUALITY】
━━━━━━━━━━━━━━━━━━

premium character bible
professional character reference photography
high-end casting sheet
AAA game character identity sheet
cinematic production reference
clean studio photography
neutral soft studio lighting
high facial fidelity
accurate anatomy
realistic skin texture
precise costume materials
consistent character identity
minimal editorial graphic design
pure white / soft neutral background
sharp details
natural proportions
high resolution

No environmental background.
No dramatic scenery.
No action pose.
No exaggerated perspective.
No fisheye.
No cinematic colored lighting.
No additional characters.
No duplicate faces.
No alternative outfits.
No random accessories.
No facial expression sheet.
No text bubbles.
No concept art sketches.
No blueprint.
No sci-fi HUD.
No decorative interface.
No character redesign.

[Laraib Fatima‎ @AiwithLariab]

GPT image 2 on chatgpt

Prompt

Transform the image into a minimalist hand-drawn doodle illustration. Use simple brush-pen outlines with slightly wobbly, imperfect strokes. Reduce all details to cute, childlike shapes while preserving the main subject and composition. Naive sketchbook aesthetic, playful and whimsical character design, clean white background, minimal linework, expressive simplicity, handmade doodle style, water color ink drawing, charming imperfections, simple cartoon illustration.

[𝐌 @Strength04_X]

Nano Banana 2 × Gemini Omni flash

Building a Tiny Luxury Farmhouse with a Swimming Pool

Prompt & workflow 👇🏻

[Umesh @umesh_ai]

Prompt : Create a 30-second short film in the same handcrafted miniature papercraft style as the reference image: folded-paper characters, visible fibers, tiny imperfections, soft natural light, shallow depth of field, gentle stop-motion movement, and cozy miniature environments. Keep the pace calm, curious, and magical. Use dialogue sparingly. Let natural ambience carry most scenes, with a soft piano and glockenspiel theme appearing only during discoveries and emotional moments. 15 shots: Morning. The boy ties his shoes beside the front door while rain taps the window. His mother hands him a tiny paper umbrella. “Looks like rain.” Outside, he notices a little red balloon caught in a tree. A handwritten tag hangs from it: “Please bring me home.” Curious piano notes begin as he frees the balloon. He walks through a misty vegetable market, balloon floating behind him. An old fruit seller points uphill. “I saw one like that yesterday.” The boy cycles toward a quiet hill, following tiny red ribbons tied to fences. He stops at a railway crossing as a miniature train slowly passes. On the other side, a curious dog begins following him. Rain suddenly pours. The boy and dog shelter inside an abandoned bus stop. He notices a child’s drawing taped beneath the roof: the same red balloon. Following the final ribbon, they reach a tiny house beside a sunflower field. A little girl opens the gate and gasps, “You found Ruby!” The music gently returns as she hugs the balloon. The boy rides home at sunset with the dog now sitting proudly in his bicycle basket. He laughs, “I guess you’re coming too.”

[Arvin @Arvin010717]

视频制作提示词：

输入参考图：故事板 【参考优先级必须锁死】 ，【@图片1】是本视频全部内容的唯一执行标准。 视频中的分镜顺序、时间轴、动作内容、攻防关系、人物站位、距离变化、运动方向、武器轨迹、法术轨迹、镜头运动、景别切换、空间位移、场景破坏、特效爆发和结尾状态，全部严格按照【@图片1】逐格执行。 禁止模型自行补充、删减、改编、重排或替换故事板内容。 【执行规则】 严格按照【@图片1】01—15格顺序生成，逐格动画化。 每格对应【@图片1】标注时间推进，禁止跳格、并格、乱序、提前、拖延、重复。 每次切镜必须承接上一格动作的方向、惯性和受力，不能断招，不能重起动作。 【@图片1】中所有角色动作、法术释放、武器攻击、机位切换、景别变化、环境互动，都必须严格服从【@图片1】画面内容，不得额外创造新招式、新镜头、新场面。 【箭头与轨迹规则】 故事板中的箭头、虚线、轨迹线、方向线、示意线，只用于控制生成时的人物运动、武器运动、法术运动和镜头运动。 它们是幕后控制信息，不是成片画面内容。 最终视频画面中绝对不能出现任何箭头、虚线、轨迹线、运动标记、示意线、编号、时间码、文字说明、分镜边框。 【画面要求】 15秒，16:9横版，720P高清，纯二维中国水墨武侠动画。 高燃、高速、狠、快、爆，动作连续不断。 整体为高完成度二维水墨打斗美学，人物清晰，动作清晰，水墨跟随强烈，泼墨爆发明显。 不要3D，不要真人，不要CG感，不要素描，不要静态插画，不要游戏感。 【风格锁定】 寒汐的效果统一为：青蓝水墨、冷白飞白、水环、水壁、水流、水纹、结界、浪潮。 赤刃的效果统一为：赤红火墨、黑色墨浪、实体红刀刀势、金色碎墨。 所有特效的规模、时机、方向和爆发关系，全部以故事板为唯一依据，不得自行夸张或删减。 【场景锁定】 场景严格按【@图片1】执行，为雨后破碎水庭废墟。 必须保留大面积浅水青石地、水面倒影、水花、水纹、断桥、残柱、倒塌水榭、碎瓦、石阶和远处墨云。 环境破坏、水面波动、碎石飞散、水壁破裂、青石开裂等，全部按照故事板指定节点发生。 【镜头锁定】 镜头语言严格服从【@图片1】。 允许高速跟拍、贴地低机位、全身中景、武器特写、俯拍、仰拍和爆撞镜头，但一切出现顺序和内容都必须以故事板为唯一参考。 切镜必须服务于故事板，不能脱离故事板自由发挥。 禁止无意义乱晃，禁止丢失主体，禁止镜头方向错乱。 【动作原则】 全程绝对实时高速战斗。 禁止慢动作、禁止快慢帧、禁止速度坡变、禁止停帧、禁止摆拍、禁止碰撞后静止、禁止一招一停。 从第一帧打到最后一帧，持续攻防，持续位移，持续受力反馈。 人物必须有重量、惯性、反作用力和明确打击感。 【声音要求】 只保留战斗拟音与环境物理声： 踏水声、水花声、水浪声、水壁破裂声、刀刃破风声、大刀劈水声、金属震响、披风破风声、碎瓦飞散声、残柱断裂声、墨浪爆裂声、滑退声、废墟回响。 不要配乐，不要鼓点，不要人声，不要对白，不要吟唱，不要动物叫声。 【负面限制】 禁止脱离故事板自由创作 禁止新增或删减故事板内容 禁止错误理解人物关系 禁止人物变脸、换装、改发型、改武器 禁止寒汐站桩施法 禁止赤刃远距离空挥 禁止红刀变成光剑 禁止水壁变成玻璃墙或普通光盾 禁止特效稀少 禁止背景空白 禁止小空间原地互殴 禁止人物轻飘、武器无重量、打击无反馈 禁止箭头、轨迹线、字幕、文字、编号、logo、水印出现在最终画面 最终原则： 人物长相、服装、武器、法术视觉听人物设计图； 除此之外，所有视频内容一律只听【@图片1】。 最终成片必须像“故事板被逐格严格动画化”，而不是根据文字重新创作。

[南鸢 nuyoah @nanyuan0412]

https://t.co/xgaDzQoGTx