AI偶像 化妝間 對話提示詞 使用工具:豆包 模型: seedance 2.0 攝影機：現代全片幅數位電影攝影機，進行手持 POV 拍攝，偶爾放置在化妝台上。自然的晃動感、細微的鏡頭震動、不完美的構圖、偶爾出現的延遲自動對焦、輕柔的變焦調整，臉部偶爾會被裁切。攝影機本身不會入鏡。 視覺風格：乾淨、高階的數位影像，搭配柔和且修飾膚色的化妝台燈光。化妝台燈泡產生細微的光暈效果，真實的皮膚紋理、淺景深、電影級動態範圍、自然色彩，具備頂級無反光鏡相機的畫質。 風格：休閒美妝 Vlog。節奏輕鬆、閒聊感，步調緩慢，帶有安靜的鏡中時刻。 參考圖片1，20 歲出頭的韓國女性偶像，留著一頭烏黑亮麗的長直髮，中分自然髮線，膚質白皙水潤，棕色雙眸，自然妝容搭配珊瑚色亮面唇彩。她穿著華麗的黑色 K-pop 舞台服裝，飾有水晶與金屬細節，肩上披著專業的黑色化妝斗篷，僅在領口、袖口與下擺隱約可見閃亮的舞台服裝。看不見的造型師雙手偶爾會進入畫面，進行梳理、分區、捲髮與造型。 場景：演出前專業的 K-pop 偶像後台化妝間，設有大型燈泡化妝鏡、造型椅、吹風機、捲髮棒、梳子、造型噴霧、化妝品，以及掛滿舞台服裝的衣架、鏡子，呈現真實的後台氛圍。 分鏡腳本： 攝影機放置在化妝台上。 親切地對著鏡頭微笑並揮手。 台灣甜美口音：「大家好！今天來做頭髮和化妝。」 手持特寫鏡頭，造型師輕柔地分開並拉動 圖片1的頭髮，準備進行造型。 台灣甜美口音：「準備過程總是需要花很長時間。」 插入造型師使用專業工具夾住並整理頭髮的特寫鏡頭。 僅有後台沙龍的環境音。 攝影機再次放置在化妝台上。 看著鏡子，露出自然的微笑。 台灣甜美口音：「現在快完成了。」 手持特寫鏡頭， 透過鏡子安靜地看著造型師，在進行最後修飾時露出溫柔的微笑。 僅有環境音。 手持鏡頭， 使用手持鏡從多個角度檢查髮型背面。 台灣甜美口音：「我也看一下後面。」 緊湊特寫， 在看到完成的髮型後，表情變得明亮起來。 台灣甜美口音：「我非常喜歡！」 自拍鏡頭。比出大拇指，熱情地對著鏡頭揮手，露出燦爛的笑容。 台灣甜美口音：「完成！下次再見。掰掰！」

神明模板提示詞 可以用我的 gam 全家便利商店海報設計 https://gemini.google.com/....../1xlPukFpTJtFcAZ5KmYlF...... 7-11海報設計 https://gemini.google.com/....../15qdfJY9OuIHponYfDeKT...... ------------------------- 給他神明[名稱] 自動生成提示詞給你 再貼上 chat gpt生成圖片 --------------------------------------- 或者參考下面提示詞 Prompt: 7-ELEVEN 版「土地公」，Q版可愛插畫風格。土地公穿著融合 7-11 經典紅橘綠條紋制服的改良神袍，手持金色拐杖，拐杖頂端掛著一個迷你御飯糰。另一手拿著 CITY CAFE 咖啡杯。土地公慈祥微笑，坐在由思樂冰杯與茶葉蛋組成的神座上。腦後光環是 7-11 標誌性的霓虹燈，寫著 "Always Open"。背景是整齊的 7-11 貨架與 ibon 機台。畫面散佈「OPEN Point +1」、「好運+1」的小圖示。中央有繁體中文 POP 字體：「無所謂，24小時 7-ELEVEN 土地公都保佑我。」。整體色彩鮮艷明亮，具備現代台灣超商設計感，商業插畫品質。 Prompt: Cute Chibi Guan Yu (Guansheng Dijun) styled as a FamilyMart store manager, 24H convenience store theme, pop art poster. [Character]: Cute Chibi Guan Yu with a charming red face and long neat black beard, wearing a vibrant hybrid outfit combining a green Chinese warrior robe with a FamilyMart store manager uniform (forest green, bright blue, and white). Crown features the FamilyMart two-bar logo. He heroically holds a Green Dragon Crescent Blade mixed with a barcode scanner in one hand, and a FamilyMart member magazine in the other. [Base/Seat]: Sitting majestically on a throne made of Let's Café coffee bean bags and warm sweet potato displays. [Background]: Modern bright FamilyMart store background, featuring refrigerated shelves, coffee bar, and LED neon halo behind his head. The halo glows cyan and blue with neon text "24H", "FamilyMart", "Welcome", and "OPEN". [Typography & Layout]: Taiwan POP poster layout. Center bold Traditional Chinese text: "無所謂，24小時全家關聖帝君都保佑我。". Speech bubble top left: "Ding-Dong~ Welcome!". Sticker top right: "晦氣退散！". Floating cute text: "功德+1", "點數+1". Right yellow POP text: "義氣相挺！要加熱嗎？". [Style]: 3D render style, C4D style, vivid neon colors, vector art icons floating around (coffee cups, ice cream, rice balls, store receipts), clean Taiwan convenience store aesthetics, extremely detailed, highly crisp typography in Traditional Chinese --ar 9:16 --v 6.0

剛好逛到一則 頁面未顯示觀看數的提示詞， 整理成 WAI 工具可以直接使用的格式， 分享給大家，可以玩玩看： 這條很適合拆解「尺度轉場如何把一支普通箭矢變成完整世界」：先用大景建立戰場規模，再讓鏡頭鎖定箭矢，最後把木紋、刻痕和羽毛放大成微縮文明。真正有價值的骨架是「遠景建立規模 → 單一物件鎖定 → 連續追蹤 → 微觀世界展開」，可以替換到產品材質展示、珠寶廣告、建築縮放、遊戲預告或品牌故事的視覺轉場。 WAI 工具試試： https://waigo.ai/share/DFwijpWQ ------ Segment 1: Lone Archer Battlefield GOAL: Create a 15-second 16:9 cinematic scale-transition sequence that begins on a vast ancient battlefield, follows one arrow in extreme slow motion, and transforms the arrow surface into a detailed miniature landscape without graphic gore. CHARACTER REFERENCE No external image reference is attached. Create one original adult archer with weathered bronze armor, dirt-streaked face, controlled posture, and consistent body proportions. Maintain the same archer identity and outfit for every visible archer shot. Do not imitate any real person or copyrighted character. CHARACTER ARCHER One original adult archer in weathered bronze armor, face streaked with dirt, grounded stance, deliberate draw strength, realistic fingers on the bowstring, practical longbow handling, and restrained facial expression. No identity drift, no face change, no duplicate archer, no fantasy glow. The archer is the only foreground human character. CHARACTER ARROW One wooden arrow with a sharp metal tip, carved runes, weathered grain, dark fletching, and a single continuous trajectory from bow release to macro landscape. Keep one arrow only. No arrow duplication, no changing shaft design, no impossible flight path. CHARACTER COUNT LOCK Exactly 1 foreground human character and 1 focal arrow. Battlefield warriors and miniature-scale figures are non-distinct background silhouettes only; they must never become new foreground characters, speak, duplicate, or pull focus from the archer and arrow. ENVIRONMENT Two explicitly defined visual scales connected by one continuous arrow transition. Scale A: a vast ancient battlefield at golden hour, hazy amber sky, smoke, ash, distant banners, swords and shields, dry earth, and thousands of soft-focus background silhouettes. Scale B: the arrow surface becomes a miniature canyon landscape with carved rune walls, tiny background warriors, miniature catapults, dust fragments, flickering torches, and small banners. No other locations, no arbitrary scene changes, no readable text, no logos.

剛好逛到一則 頁面未顯示觀看數的提示詞， 整理成 WAI 工具可以直接使用的格式， 分享給大家，可以玩玩看： 這條很適合拆解「如何用一個精準轉場翻轉觀眾對場景的理解」：前半段先用連續手持鏡頭建立生存危機，再用完全對齊的手部動作做 match cut，最後才揭露 VR 的日常真相。真正有價值的骨架是「高壓情境 → 明確中斷 → 視覺對位 → 現實反轉」，可以替換到遊戲實況、沉浸式展覽、產品開箱、健身訓練或科幻廣告中。 WAI 工具試試： https://waigo.ai/share/DFwijpWQ ------ Segment 1: Polar Bear VR Twist GOAL: Create a 15-second grounded cinematic survival sequence that begins as a polar-bear attack in a frozen wilderness and resolves through one precise match cut into the same woman's ordinary living room, revealing the first-person reality behind the scene. CHARACTER REFERENCE No external image reference is attached. Create one original adult woman and maintain the same face, dark hair, body proportions, hand size, and clothing continuity through the wilderness portion. After the match cut, preserve the same face, hand height, head angle, and camera distance while changing only into loose indoor clothes and adding the VR headset at the specified time. Do not imitate any real person or copyrighted character. CHARACTER PROTAGONIST One original adult woman, athletic but realistic build, weathered fur-and-leather winter outfit during the frozen wilderness portion, later loose neutral home clothes. Controlled survival movement, anatomically correct hands, clear foot placement, grounded balance, tired practical reaction rather than exaggerated fear. She is the only human character. No identity drift, no face change, no duplicate body. CHARACTER POLAR_BEAR One massive realistic polar bear with wet white fur, heavy body weight, visible breath, and natural animal biomechanics. Fast but physically plausible leap, readable paw placement, no fantasy anatomy, no extra heads, no duplication, no graphic injury. CHARACTER COUNT LOCK Exactly 2 active subjects total: 1 original human protagonist and 1 polar bear. One off-screen voice may be heard but must never appear as a visible character. No additional humans, animals, clones, reflections, or crowd members.

剛好逛到一則 頁面未顯示觀看數的提示詞， 整理成 WAI 工具可以直接使用的格式， 分享給大家，可以玩玩看： 這條很適合拆解「召喚、變身和戰鬥預備」如何在 15 秒內交代清楚：真正有價值的骨架不是堆特效，而是把手勢、能量來源、角色狀態、服裝變化、武器顯現和最後姿勢排成一條可追蹤的因果鏈。這個結構可以替換成產品啟動、角色覺醒、品牌視覺轉場、遊戲角色展示或運動員賽前準備；只要保留「動作觸發 → 形體變化 → 能力具象化 → 清楚定格」的節奏，短片就會更有記憶點。 WAI 工具試試： https://waigo.ai/share/DFwijpWQ ------ Segment 1: Sakuga Sorcery Battle GOAL: Create a 15-second 16:9 high-quality anime transformation action sequence in which one original shrine sorceress summons a giant Japanese giant salamander spirit, fuses with it, manifests a short dagger, and ends in a poised first-strike stance. Keep the motion fluid, readable, and free of gore or injury. CHARACTER REFERENCE No external image reference is attached. Create one original adult shrine sorceress with a consistent face, hairstyle, body proportions, and outfit silhouette throughout the entire sequence. Do not imitate any real person or existing copyrighted character. Maintain identity consistency across every shot. CHARACTER sorceress Original adult female shrine sorceress, calm focused face, dark hair initially tied in a simple style, clean shrine-inspired uniform with a short skirt and practical boots. During the transformation, add subtle miko markings, a fitted combat shrine outfit, a secure waist sash, and a tightened twin-tail hairstyle. Natural anatomy, grounded human speed, controlled hand seals, precise footwork, no supernatural body distortion. Keep the same face structure, skin tone, hair color, and body proportions across all shots. CHARACTER salamander_spirit Exactly one giant Japanese giant salamander spirit familiar. Rounded head, wrinkled dark green skin, long tail, translucent ink-green and gold-violet spiritual aura, slow deliberate movement, clearly non-human creature anatomy. The spirit remains one continuous entity, never duplicates, never becomes a second human, and never creates extra faces or limbs. CHARACTER COUNT LOCK Exactly 2 entities total: 1 human sorceress and 1 salamander spirit familiar. No additional humans, creatures, clones, duplicates, background characters, crowd, or human-like reflection. Strict count maintained for the entire 15-second sequence.