提示庫（103 條）
所有模型的技巧、模板與官方英文範例提示。用上方搜尋或下面篩選。
圖像模型 GPT Image 2.5｜Flare／Sunburst 變體選擇 + 品質分級
官方 CLI 範例（產生一張橫向產品照、2K）。草稿階段用 Flare + 低品質大量試；確定方向後改 Sunburst 及 high 以上品質定稿。品質與變體是兩個獨立旋鈕。
模板
higgsfield generate create gpt_image_2_5 --prompt "<subject, setting, lighting, composition>" --variant flare --quality medium --resolution 2k --wait
官方範例提示（英文原文）
higgsfield generate create gpt_image_2_5 --prompt "A landscape product photo" --resolution 2k --wait
圖像模型 GPT Image 2.5｜資訊圖／設計稿（圖中文字的官方預設用法）
官方 CLI README 的 GPT Image 2.5 範例（譯：乾淨的全球能源結構資訊圖，扁平圖示，低飽和色板），3:4、high 品質、2K。設計類提示寫「版式類型＋內容＋圖示風格＋色板」即可。
官方範例提示（英文原文）
higgsfield generate create gpt_image_2_5 \ --prompt "clean infographic showing global energy mix, flat icons, muted palette" \ --aspect_ratio 3:4 \ --quality high --resolution 2k \ --wait
圖像模型 GPT Image 2.5｜參考圖編輯：只描述「變更」，不要重寫原圖
官方 Skills「Image-to-image」規則的 Good 範例（譯：轉為動漫風格，鮮艷色彩，柔和賽璐珞陰影）。帶 --image 時，模型已看到原圖，重新描述原圖反而令它混淆；GPT Image 2.5 的強項正是「只改被要求的地方」。
模板
<describe only what changes>, keep <subject/lighting/framing> unchanged
官方範例提示（英文原文）
transform into anime style, vibrant colors, soft cel shading
圖像模型 GPT Image 2｜用 product-photoshoot 技能的 10 種模式，而非手寫產品提示
官方 Skill 的範例意圖描述（譯：陽光廚房枱面上的一瓶冷萃咖啡，IG 動態用）。只需寫簡短意圖，技能會套用該模式的專屬模板。模式：product_shot（純色/棚拍）、lifestyle_scene（生活場景）、closeup_product_with_person（手部/半臉特寫）、moodboard_pin（2:3 Pinterest）、hero_banner（網站橫幅）、social_carousel（3–10 張連貫輪播）、ad_creative_pack（廣告組合）、virtual_model_tryout（AI 模特試用）、conceptual_product（超現實/懸浮/飛濺）、restyle（改風格/季節）。
模板
higgsfield product-photoshoot create \ --mode <mode> \ --prompt "<short user-intent description from interview answers>" \ [--image <path-or-upload-id>]... \ [--count <1-10>] \ [--aspect_ratio <override>]
官方範例提示（英文原文）
bottle of cold-brew on a sunlit kitchen counter, IG feed
圖像模型 GPT Image 2｜App UI／落地頁（準確 UI 文字），作後續影片的設計系統參考
官方 CUE 案例：整套 App 介面由 GPT Image 2.0 生成（官方說明其他模型會把 UI 內文字弄亂）。之後的畫面以「@image_1 — visual style and design system reference」保持字體、顏色一致，再放進 Marketing Studio 影片的手機畫面中。
官方範例提示（英文原文）
Modern dark minimal landing page for "CUE" — AI-powered recipe generator. Style: premium, minimal, dark green aesthetic, clean SaaS, Apple-level design Layout: Centered hero: - Clean input box (ChatGPT-style) - Placeholder: "Type ingredients or upload a photo..." - Upload icon inside input - Minimal generate button Top: - Small logo (avocado inside scan frame) - "CUE" wordmark Headline: "Cook with what you have" Subheadline: "Turn your ingredients into real meals with AI" Right side visual: - Large transparent ice cube floating in space - Inside the ice cube: frozen ingredients (tomato, egg, greens, cheese) - Ingredients slightly suspended, visible through ice - Realistic ice texture: cracks, air bubbles, frost edges - Soft light reflections and subtle glow - Clean composition, no clutter Background: - Deep dark green to black gradient - Subtle vignette Below input: - Small example text: "eggs + tomato → omelette" Design: - lots of whitespace - soft glow on input - modern sans-serif typography (Inter / SF Pro) - minimal, refined Mood: calm, premium, slightly futuristic, unique visual metaphor (frozen ingredients → potential meal)
圖像模型 GPT Image 2｜以第一張 UI 作設計系統參考，生成其他頁面
官方 CUE 案例第二步：把落地頁作 @image_1，要求「Same design system as @image_1」生成食譜結果頁。
官方範例提示（英文原文）
@image_1 — visual style and design system reference for CUE app. Same design system as @image_1. Recipe result screen — title, cook time, ingredients, step-by-step instructions, food photo. Keep the same dark premium aesthetic.
圖像模型 GPT Image 2｜CLI 基本生成
官方最短範例（譯：一張棚拍產品照）。GPT Image 2 預設 quality=high、2k。
官方範例提示（英文原文）
higgsfield generate create gpt_image_2 --prompt "A studio product photo" --wait
圖像模型 Nano Banana Pro｜精確數量＋排列約束
官方專家用例：要求 20 人按 7/6/7 三排排列並各持獨特物件——測試模型對數量與佈局的服從度。中文要點：先講總數，再用括號列明每排人數，最後描述每人的差異。
官方範例提示（英文原文）
An image of an office where 20 different people are neatly arranged (7 in the front row, 6 in the middle row, 7 in the back row), each holding a unique object whose name starts with the letter 'S'.
圖像模型 Nano Banana Pro｜技術資訊圖／拆解圖（借用已知視覺語言）
官方用例：F-117 夜鷹的資訊圖拆解。關鍵是「similar to Formula 1 car schematics」這類**風格類比**，一句就定下整體排版語言。
官方範例提示（英文原文）
High-resolution infographic breakdown illustration of a F-117 Nighthawk. Crisp, detailed technical diagram style similar to Formula 1 car schematics. White handwritten-style arrows and labels pointing to all key components. Clean engineering aesthetic, a bit like automotive blueprint annotations. Full object visible in dynamic perspective, sharp lighting, high clarity. Background clean and slightly blurred for emphasis on the subject. Add multiple callouts around the object with lines and text, e.g.: ‘Component Name’, ‘Function / Purpose’, ‘Material Type’, ‘Dimensions’, ‘Power / Capacity / Performance Stats’. Use consistent white annotation lines, diagram boxes, arrows, and outlines. Infographic title at the top in a clean modern font: ‘[F-117_Nighthawk] - Technical Breakdown.’
圖像模型 Nano Banana Pro｜漸變序列（同一物件的多個狀態）
官方用例：七塊牛扒由生到全熟排成一列。適合教學圖、比較圖：寫明數量＋排列方向＋每個狀態的差異。
官方範例提示（英文原文）
A high-resolution food photograph shows seven cuts of steak, sliced and arranged in a row on a wooden board, displaying the full gradient of doneness from Blue Rare to Well Done, set in a modern kitchen.
圖像模型 Nano Banana Pro｜高密度場景（50 件物件）
官方用例：時間旅行者書房內 50 件真實歷史文物（原文其後逐項列出 50 件文物及房間環境；部分文物名含商標，本庫只節錄開首段）。展示 NB Pro 的世界知識與高密度細節能力。
官方範例提示（英文原文）
Create an ultra-realistic, richly textured image of an eccentric time-traveler’s private study, filled with 50 real historical artifacts, displayed in perfect clarity. The room is dimly lit with warm tungsten lamps, soft shadows, dust floating in the air, wooden furniture, brass mechanisms, and worn leather textures. A large panoramic desk occupies the center, surrounded by shelves, crates, cabinets, glass domes, and wall mounts. Every item below must be clearly visible, physically placed, never floating, arranged naturally on shelves, the desk, the floor, or inside cases.
圖像模型 Nano Banana Pro｜極短提示也能成立的場景
官方用例（譯：拍攝花絮廣角，劇組正在佈置電影感人像場景）。官方說明此例展示「即使簡單提示也能控制場景構圖與細節」。先用短提示探索，再逐步加約束。
官方範例提示（英文原文）
Behind-the-scenes wide shot showing a film crew setting up a cinematic portrait scene.
圖像模型 Nano Banana Pro｜CLI 範例：主體＋材質＋光線（三段式短提示）
官方 CLI README 的 Nano Banana Pro 範例（譯：現代建築，玻璃幕牆，黃金時段光線），16:9、2K。注意 ID 是 nano_banana_2（= Pro）。
官方範例提示（英文原文）
higgsfield generate create nano_banana_2 \ --prompt "modern architecture, glass facade, golden hour light" \ --aspect_ratio 16:9 \ --resolution 2k \ --wait
圖像模型 Nano Banana Pro｜參考圖補光編輯（保留表情）
社群用戶分享（譯：給他棚拍燈光，但不要改變他的表情）。示範「改什麼＋保留什麼」的一句式編輯。
官方範例提示（英文原文）
Give him studio lighting without changing his facial expression.
圖像模型 Nano Banana 2（Gemini 3.1 Flash Image）｜多層約束提示（物件＋排列＋光線＋字體＋資訊層）
官方網誌示範「分層提示」（譯：三件物件對稱排列的平鋪產品照，暖光，手寫字體，可見資訊圖覆蓋層）。官方指出一般模型難以處理這類多層要求，而 NB2 可一次滿足。
官方範例提示（英文原文）
Create a flat lay product shot with three objects arranged symmetrically, warm lighting, handwritten typography, and a visible infographic overlay.
圖像模型 Nano Banana 2（Gemini 3.1 Flash Image）｜角色／卡通風格＋參考圖
官方 Skill 範例（譯：動漫角色概念，誇張姿勢），配 --image ./ref.png。卡通／插畫角色是 NB2 的自動選擇場景。
官方範例提示（英文原文）
anime character concept, expressive pose
圖像模型 Nano Banana 2 Lite｜快速參考圖編輯
官方 CLI 範例（譯：乾淨的產品草圖，白色背景）。
官方範例提示（英文原文）
higgsfield generate create nano_banana_2_lite --prompt "clean product sketch, white background" --image ./reference.png --resolution 1k --wait
圖像模型 Higgsfield Soul 2.0｜以 Soul ID 生成一致角色（CLI）
模板為官方 Soul ID Skill 原文；範例提示來自官方 CLI README 的 Soul 範例（譯：專業肖像，中性背景，柔和日光）。--soul-id 會以 custom_reference_id 送出。
模板
higgsfield generate create text2image_soul_v2 --prompt "..." --soul-id <ref_id> --quality 2k --wait
官方範例提示（英文原文）
professional portrait, neutral background, soft daylight
圖像模型 Higgsfield Soul 2.0｜自建 Moodboard（自訂風格預設）
四步：①收集最多 80 張風格一致的圖（約 10 張已足夠，20–30 張更好）；②上載到 Moodboards；③Soul 2.0 學習並建立預設；④生成時套用。避免低質、模糊、混合風格及人臉。
圖像模型 Higgsfield Soul 2.0｜Soul HEX 色彩控制
文字描述顏色很難準（「warm gold」可能變芥末黃）。Soul HEX 讓你從參考照片抽取色板或直接用 HEX 色碼（#RRGGBB），生成符合品牌指引的配色。
圖像模型 Soul Cinema（soul_cinematic / soul_cinema_studio）｜Soul ID 角色的電影感劇照
官方 Skill 指令。Skills 自動選擇規則：已訓練的 Soul 角色要「電影感」→ soul_cinematic；一般電影劇照也選它。
模板
higgsfield generate create soul_cinematic --prompt "..." --soul-id <ref_id> --quality 2k --wait
圖像模型 Soul Cast｜參數化「固定旁白角色」
官方用例：電影敘事頻道要一位固定旁白——設定 Sage 原型、年長女性、銀髮、正式服裝、60 多歲，一次設定、之後重用。重點：Soul Cast 的「提示」其實是參數組合，文字提示為選填。
圖像模型 Soul Location｜用正面描述取代「no people」
官方 Skills 負面詞規則：不要寫「no people」，改寫「uninhabited landscape」（無人之境）。Soul Location 本身就是無人場景模型，提示專注於地點、時間、天氣、材質。
模板
[location type], [time of day], [weather/atmosphere], [materials & textures], [camera/lens], uninhabited
官方範例提示（英文原文）
uninhabited landscape
圖像模型 Seedream 5.0 Lite｜用自然語言描述意圖，而非堆砌關鍵字
依官方評測歸納：Seedream 5.0 Lite 會評估「提示背後的意圖」，所以把用途、情緒、物件之間的位置關係用一句完整的話講清楚，比逗號分隔的關鍵字更有效。涉及近期事件時可受惠於即時搜尋（注意：涉及真實人物仍受平台安全規則限制）。
模板
[What the image is for] + [mood/atmosphere] + [spatial relationships] + [the change you want, in plain language]
圖像模型 Flux Kontext Max｜風格轉換：只寫目標風格
官方 Skills media-inputs 的 --image 編輯範例（譯：轉為水彩風格）。對 Kontext 類編輯模型，提示只寫「要變成什麼」。
官方範例提示（英文原文）
stylize in watercolor
圖像模型 Recraft V4.1（含 Styles / Utility）｜向量 Logo（CLI）
官方 CLI 範例（譯：咖啡品牌的極簡向量標誌）。寫「風格（minimal）＋類型（vector logo mark）＋用途（for a coffee brand）」，並以 --model_type vector 取得向量輸出。
官方範例提示（英文原文）
higgsfield generate create recraft_v4_1 --prompt "minimal vector logo mark for a coffee brand" --model_type vector --resolution 2k --wait
圖像模型 Z-Image｜「鏡頭類型：」標籤開頭的短段落提示
官方範例：先用粗體標籤定鏡頭類型（Hyper-realistic wildlife close-up），再寫主體、材質（濕毛）、眼神、背景色。短句、具體感官細節。
官方範例提示（英文原文）
Hyper-realistic wildlife close-up: Tiger head emerging from bright green duckweed. Wet fur, intense amber eyes focused on the camera. Smooth vibrant green water background contrast. Sharp details, soft natural light.
圖像模型 Z-Image｜電影感人像（光源顏色＋方向）
官方範例：寫明光源（綠與紅霓虹背光）與場景（昏暗地下空間），括號補充造型細節。
官方範例提示（英文原文）
Cinematic portrait: Young woman (windswept dark hair) leans on a railing in a dim underground space. Lit by moody green and red neon backlighting, creating dynamic motion. Confident, mysterious expression. High realism, shallow depth of field, futuristic urban photography.
影片模型 Seedance 2.5｜戲劇外景：人物＋自然光（含事件時間軸、PHYSICS、TEXTURE）
官方 10 類測試之一（戲劇）。重點：ACTION TIMING 精確到秒（浪在 1.2s 爆開、風在 2.4s 吹起頭髮、5.4s 說出台詞）；PHYSICS 描述浪、頭髮、眼淚的物理；LIGHTING 只有「一個真實光源：太陽」；AUDIO 明言「No music, no narration, no subtitles」；TEXTURE 指定 IMAX 65mm 菲林與「No beauty retouch」。
官方範例提示（英文原文）
Golden hour on a cold rocky shore. A young man and a young woman sit side by side wrapped in wool blankets, facing the low sun over the sea, both with tears standing in their eyes. She turns and tells him she will always love him; he looks down, looks back, smiles softly and sadly, answers, and they both return their gaze to the sea. Waves break on the rocks below them the whole time. CHARACTERS: THE MAN, late 20s, lean build, weathered light skin flushed at the nose and cheekbones from the cold wind, short dark brown hair blown messy, light stubble, grey-green eyes with heavy lower lids, a strong straight nose. Wrapped shoulders-to-hips in a heavy charcoal wool blanket with a frayed edge, a cream cable-knit sweater collar showing at his neck, dark trousers, worn leather boots braced on the rock. Hands hidden inside the blanket. No logos, no readable text on anything. A wholly original, invented face. THE WOMAN, always on the man's LEFT side: mid-20s, slight build, long chestnut wavy hair whipping loose in the wind, light freckles across the nose, hazel eyes, lips chapped pale from the cold. Wrapped shoulders-to-hips in an oatmeal wool blanket with a woven fringe, a rust-colored knit sweater collar showing, dark jeans, ankle boots. Hands hidden inside the blanket. No logos, no readable text. A wholly original, invented face. Only these two people exist in the entire video. No other figures, no boats, no birds close to camera, no animals on the rocks. LOCATION: A rocky northern shoreline at golden hour, the sun four degrees above the sea horizon in the west, minutes from setting. The couple sits shoulder to shoulder on a flat dark rock shelf two meters above the waterline, facing west, directly into the sun. Below and ahead of them, five to eight meters out, a scattered line of wet black boulders takes the breaking waves. The rock is dark basalt grey, wet and glistening on its seaward faces, dry and matte where they sit, with pale barnacle crusts at the waterline and pockets of coarse sand between stones. Cold onshore wind blows from the sea into their faces. The sky is clear amber gold near the sun grading to pale cool blue overhead, the sea is dark steel blue scattered with thousands of golden glints. No buildings, no people, no signage, nothing readable anywhere. SEA AND WIND EVENT TRACK: The sea is an event track, not a texture. Wave impacts arrive at irregular intervals, no two identical. 0.0s, mid swell rolling in, leftover foam draining through the channels between rocks. 1.2s, WAVE 1 strikes the forward boulders, bursting spray to 12 percent of frame height, the cloud drifting inland toward the couple on the wind. 2.4s, GUST 1, hair and the free edges of both blankets lift and snap once, then settle. 3.9s, WAVE 2, smaller, striking the offset rocks screen left, 6 percent. 5.0s, WAVE 3, 8 percent, its hiss fading out just before she speaks, a relative lull holds under her line. 7.6s, GUST 2, the strongest of the shot, both bodies brace a degree into it. 9.6s, WAVE 4, 10 percent, its backlit spray drifting as golden mist through the background of his close-up. 12.4s, WAVE 5, 7 percent. 13.8s, WAVE 6, the largest, 14 percent, a low thud felt through the rock, tall golden spray hanging a beat in the backlight. FORMAT MODE: Controlled five-segment sequence, four HARD CUTs, 15.0 seconds total. Real-time motion. Continuous ambient sound across all cuts. Exactly two scripted spoken lines, nothing else spoken. No music, no narration, no subtitles. STILLNESS LOCK: Both remain seated on the rock shelf for the entire video. Nobody stands, nobody hugs, nobody kisses, nobody's hands leave the blankets. Both keep tears standing in their eyes for the whole film, exactly one tear falls, hers, at 13.2s. FIRST FRAME AND SPATIAL BLOCKING, SEGMENT 1: Camera behind the couple looking west. THE WOMAN seated screen-left, x 42%, head at y 44%. THE MAN seated screen-right, x 58%, head at y 42%. Both backs to camera, both faces aimed at the horizon. Reverse-angle note for Segments 2, 3 and 4: with the camera crossed to the seaward side, THE WOMAN appears screen-RIGHT and THE MAN screen-LEFT. They never swap sides. CAPTURE FORMAT AND OPTICS: Shot on IMAX 15-perf 65mm color negative with Panavision System 65 anamorphic optics, 2x horizontal squeeze, unsqueezed in post. LENS LOCK: Segments 1, 5 at 47°, deep focus holding the couple, the breaker rocks and the sun. Segments 2, 3 at 29°, short telephoto portrait, camera 4 meters from the face, aperture wide open, thin focus with the sea dissolved into drifting golden oval bokeh. Segment 4 at 29°, two-shot, camera 5 meters back, both faces in the same focus plane. CAMERA: Segment 1, heavy IMAX body on a wheeled dolly at 1.2 meters, traveling forward 2 meters, constant slow speed. Segments 2, 3, 4, reverse angle from the seaward rocks, shoulder-mounted with large-format weight. Segment 5, same axis as segment 1, dollying slowly backward 1 meter. ACTION TIMING: 0.0s to 3.0s, SEGMENT 1, wide from behind, dolly-in. Both already seated, WAVE 1 bursts at 1.2s, GUST 1 lifts her hair at 2.4s. HARD CUT. 3.0s to 7.2s, SEGMENT 2, her close-up. At 5.4s she says, soft and unsteady: "I will always love you." HARD CUT. 7.2s to 11.6s, SEGMENT 3, his close-up. At 10.6s he says, quiet and rough: "Me too." HARD CUT. 11.6s to 13.4s, SEGMENT 4, two-shot. At 13.2s a single tear breaks from her right eye. HARD CUT. 13.4s to 15.0s, SEGMENT 5, wide from behind, pull-back. WAVE 6 thuds at 13.8s, their shoulders settle together at 14.2s, the only touch in the film. PHYSICS: Every wave follows the event track, surge, impact, burst, ballistic spray bending inland with the wind. Hair strands whip, tangle and release, never static. Tears behave physically, a trembling meniscus held at the lash line, the single falling tear at 13.2s obeys gravity and skin, bent slightly by the wind. LIGHTING: One real source, the sun, four degrees over the sea, warm deep gold, unmoving for the full 15 seconds. Against it, a cool pale blue skylight fill from above. Segments 1 and 5, from behind, the couple reads as two dark wrapped shapes rimmed in gold. Segments 2, 3, 4, frontal, full low golden key on the faces, wet eyes carry one small intense golden catchlight each. AUDIO: Continuous cold shore ambience, wave impacts exactly at the event track times, foam draining between rocks, steady wind buffet with the two gusts, one distant gull cry at 8.6s. The two scripted lines are the only words. No music, no narration, no subtitles. TEXTURE AND REALISM: IMAX 15-perf 65mm negative, fine organic grain, deep blacks with a milky filmic floor, gentle halation only around the sun. Skin unretouched and hyper-detailed, pores, chapped lips, windburn, wet lash clumps. No beauty retouch, no digital smoothing, no CGI sheen.
影片模型 Seedance 2.5｜動作多鏡頭編排
官方動作類範例。以「Style:」開頭定下「NOT a clean CGI render」的總規則，再逐段寫動作與物理質量（紙張的真實重量與慣性）。
官方範例提示（英文原文）
Style: 8K large-format photoreal action cinema, gritty, textured, photochemical, NOT a clean CGI render. A lone streetwear heroine's heist interrupted into brutal close-quarters combat in a cargo hold at altitude, fought in a storm of living paper money. References in spirit, fully original execution: Christopher Nolan's Tenet for practical altitude weight, Denis Villeneuve's Sicario for cold procedural tension, the grounded brutal hand-to-hand grammar of modern action thrillers, Michael Bay's kinetic camera energy stripped of pyrotechnics, cinematographer Hoyte van Hoytema and Greig Fraser for hard naturalistic light. Real stunt cinema, weight, impact, breath, imperfect optics. TEXTURE AND REALISM (critical, defeats the plastic look): photoreal documentary-plate realism throughout. Real optical imperfection on every frame, light authentic 35mm film grain, gentle gate weave, halation bloom on highlights, faint chromatic aberration at frame edges, sensor grain in shadows, lens breathing on focus pulls, real anamorphic flare, natural micro motion blur. Skin, metal, fabric and paper carry real-world micro-detail and subsurface light response. NOT 3D-render, NOT plastic CGI sheen, NOT glossy synthetic, NOT video-game look. Battered, lived-in, photochemical. Cinematography: Hybrid energy, implied gyro-rig for the entry, then aggressive handheld combat chaos, destabilized, reactive, whip-pans tracking strikes, fast push-ins on impact, one orbital around the clash. DP references: Hoyte van Hoytema, Greig Fraser, Roger Deakins, Wally Pfister, cold naturalistic, never glossy. Lighting: Source, a hard shaft of cold daylight through the half-open rear ramp, cold LED strips along the fuselage ribs, deep shadow in the recesses. Direction, hard back-and-side ramp light throwing both fighters into rim-lit silhouette as they cross the beam, the floating banknotes catch the light shaft. Quality, hard, high-contrast, sculpted, dramatic. NOT golden-hour, NOT soft, clinical cold interior light with the warm accent of her hair and the warm paper. Color: 60 percent cold steel-blue and gunmetal-grey (hold interior, walls), 30 percent matte black (enemy gear, cargo, jetpack harness), 10 percent warm focal pop (her coral-orange hair, white tee, white star bandana, the warm paper money). Cold procedural action grade, crushed blacks, desaturated steel mids, single warm accent. She and the cash are the only warm things in a cold world. NOT golden-hour Hollywood, NOT teal-and-orange blockbuster. Camera: Arri Alexa Mini LF plus Master Anamorphic for interior combat, Arri Alexa 65 for the entry, Laowa 24mm Probe Lens for money and impact macro inserts. Shallow-to-medium depth of field, the hold receding into haze and shadow behind the fighters. Base 24fps. The decisive counter rendered as 120fps deep slow motion played at 24fps. Light authentic 35mm film grain, subtle anamorphic flare on the ramp light. Skin: Pore-level realism preserved exactly as in the reference. HER: East Asian young woman, fair light complexion with natural flush, dark brown eyes, straight dark brows, soft full lips, high cheekbones, wind-reddened then sweat-sheened from the fight, thin metal glasses catching a cold glint, knocked askew mid-fight. ENEMY: weathered olive-skinned man, hard angular features, dark stubble, a faint old scar through one brow, sweat on the brow. Natural pore texture, no over-retouching, the non-glamour of photoreal action. Acting: HER, cold focused thief turning fierce grounded fighter, rehearsed economy as she works the case, a flash of alarm at the interruption, then committed aggression, reading her opponent, striking with full body weight. ENEMY, broad, powerful, methodical, a predator who catches her mid-theft and lunges. NOT theatrical, NOT wire-fu posing, silent, brutal, breath-driven. Her gaze stays on the cash, then on the enemy, never the lens, one final beat she looks toward the open ramp. Physics: Real fight physics, strikes connect with genuine weight, recoil and follow-through, bodies absorb impact and stagger with believable momentum, no floaty wire-fu. When she slams into the cargo the straps and nets flex and the metal booms. LIVING MONEY (critical, not plastic blocks): the currency is real paper, individual banknotes, NOT solid rigid bricks, NOT monolithic CGI blocks. Paper bills with visible fibre texture, printed ink detail, soft worn creased edges, slight curl and limpness. When the bands snap, individual notes separate and fan out, each tumbling and fluttering on its own air current in the cold draughty hold, crinkling, folding mid-air, catching light on flat faces, spinning edge-on, drifting at different speeds. The fight churns through this living storm of paper, bills kicked up by every movement, swirling around the strikes. Real paper mass, drag and inertia. Her coral-orange hair and bandana ends whip, the white tee and polka-dot trousers move with real fabric weight, cold wind gusts through the half-open ramp. Composition: Heroic large-format framing, 16:9. The hold recedes in depth into haze and shadow, the ramp light shaft cutting a hard diagonal, drifting banknotes catching the beam. The two figures cross in and out of the light, rim-lit silhouettes against the blown-out rear. Tight on the case-open and on impacts, wider on the clash amid the floating cash, low angle on the takedown. Final hero beat: she stands in the ramp light, bag in hand, bills settling around her. Continuity: Two characters throughout, HER (same face, coral-orange hair with blunt bangs, white star-print bandana, thin metal glasses, black beaded choker, white cropped tee, navy polka-dot drawstring trousers, black cross-body bag, jetpack harness, fingerless gloves) and the ENEMY (consistent matte-black tactical gear, scarred brow, stubble), kept consistent across all 9 shots. Same cold light, same steel-blue grade, same hold, same drifting cash. One continuous real-time sequence. Same generic grey turboprop cargo lifter. Editing: Kinetic grammar, 9 shots in 15 seconds, varied pacing, steady on the theft (1.2 to 1.8s), a sting cut on the enemy reveal, snap and whip-pan cuts on the strike exchanges, a held 2.0s beat on the slow-motion counter. Clean cuts and whip-pan cuts. Cuts land on beats, the bands snapping, the enemy stepping in, the lunge, the slam, the counter connecting, the body hitting the floor. No transitions, no fades. Final hero beat holds, then hard cut to black. Technical: 16:9 widescreen, cinematic.
影片模型 Seedance 2.5｜商業產品：多參考場景（手持、20 秒）
官方商業類範例：以「SCENE CONTEXT」開場，產品與人物以參考圖鎖定，手持輕晃增加真實感。
官方範例提示（英文原文）
SCENE CONTEXT: A 20-second product commercial shot entirely handheld with a light natural shake, a girl strolls through a sunny city park among everyday walkers and spots something glowing, levitating over the lawn, a pair of pink over-ear headphones floating in the air. She reaches out, touches them, and in a blink they're on her ears. She takes one step and the ground recolors under her foot in a wave of vivid gradient color while every person in the park freezes mid-motion. With each step the world repaints harder, until the final gag shot, a frozen grandma mid-stride and her little dog on the leash, very much not frozen, barking at it all. The product bends reality, the dog doesn't care. ACTIVE REFERENCES: The girl, identity and wardrobe locked to the reference, ignore the reference backdrop, young East Asian woman, early 20s, chin-length dark-brown wavy bob with soft face-framing curls, warm brown eyes, rosy blush, glossy nude-pink lips, small pearl studs and a pearl choker, a cornflower-blue fitted tee with white lace trim and three little white bows down the front, a white tiered eyelet-lace mini skirt, white cowboy boots with silver star details. Face lock: photoreal natural skin, identical features in every shot, zero drift, no smoothing. Location, as-is from the reference: the wide park lawn, deep green grass, a big dark tree frame-left, the treeline behind, and the city skyline over it, the ornate spired tower, stone mid-rises, glass supertalls, white clouds in blue sky. All signage unreadable. The product, inherited exactly from the reference: premium over-ear wireless headphones, blush-pink padded leather headband with scalloped cushions, rose-gold metal sliders and accents, two-tone ear cups, blush-pink outer shells, cream-ivory inner pads, small side buttons, unbranded, no logos. The hero object, every detail per the asset when floating, when touched, and when worn. THE WORLD-PAINT EFFECT (critical staging): The recolor spreads only from her footfalls, each step sends a visible wave of saturated gradient color rolling outward across the ground like liquid light, grass sweeping dark-green to light-green, paths flushing orange to purple, the sky and glass towers washing blue to electric blue as the waves reach them. Every wave travels with honest radial speed, repainting surfaces it crosses and staying, untouched areas keep their natural color until a wave arrives. The freeze hits all people at her first step, every walker locks mid-pose like statues, only she, the color waves, the leaves in the wind and the dog stay alive. Format mode: Five segments, hard cuts at 4.0s, 8.0s, 10.5s, 16.5s, commercial cutting. Real time inside each segment, the girl, headphones and park identical across cuts, zero drift. Optics: bright commercial-film look, crisp daylight exposure, her and the product razor-sharp, gentle depth on the park, clean saturated color science that makes the gradient waves sing. Camera: true handheld with light shake throughout. Segment 1, a walking follow beside her. Segment 2, a curious push-in behind her shoulder toward the glow. Segment 3, a tight orbit-step around the touch. Segment 4, energized backpedal ahead of her strides, dipping low to catch each footfall wave. Segment 5, a settled handheld close-up hold for the gag. Action timing: Segment 1 (0.0s to 4.0s), the walk, she strolls the lawn path in the sun, at 2.8s a glow flickers over the lawn ahead. Segment 2 (4.0s to 8.0s), the find, over-shoulder push-in as she approaches the floating headphones, wrapped in a soft rose light. Segment 3 (8.0s to 10.5s), the touch, tight on her fingertips meeting the shell at 8.6s, a soft pulse of light, and in a blink the headphones are on her head. Segment 4 (10.5s to 16.5s), the steps, she plants her right boot at 10.9s and a color wave detonates from the footfall, rolling the grass dark-green to light-green while every person freezes; second step at 12.4s, an orange-to-purple wave; third at 13.9s, the sky flushes blue to electric blue; fourth at 15.4s, the waves overlap. Segment 5 (16.5s to 20.0s), the gag, handheld close on the frozen grandma, locked mid-stride, and at the leash's end her little dog is fully alive, hopping, spinning, barking indignantly at the frozen world. Physics: real magic on honest rules, the headphones levitate with a gentle bob and slow rotation, the touch-transfer is one clean light pulse, no morphing hands. Each color wave radiates from the exact footfall point at constant speed, repainting and staying. Frozen people are absolute statues, zero sway, clothes and hair locked. The dog is fully dynamic, real bark mechanics, leash going taut against the frozen hand. Lighting: bright natural park daylight, sun from high frame-right, soft cloud fill. The levitating headphones carry their own warm rose glow that dies into them at the touch. The gradient waves add saturation, not exposure, the light stays true daylight while colors repaint. Audio: diegetic-plus-product sound, the park alive in Segment 1, birds, distant chatter, the levitation hum rising soft and warm from 2.8s, the touch pulse at 8.6s, a clean deep harmonic bloom, then the world's ambience ducks underwater-quiet as the freeze lands, leaving a warm muffled bass groove breathing from inside the headphones, her boot-steps ringing bright, each footfall firing a soft synth-wash whoosh. In the finale, the dog's sharp indignant barking cutting the hush. Positive locks: Five segments, hard cuts only. Her identity and the headphones exactly in every frame, never redesigned, no logos. The freeze hits only at her first step and holds, no frozen person ever moves again, the dog never freezes. The color waves come only from her footfalls, repaint permanently, and never strobe the exposure. Nobody speaks, only the dog barks. All text unreadable, everything unbranded.
影片模型 Seedance 2.5｜史詩風景：環境當主角（三鏡同一推軌）
三個鏡頭都用「slow steady push-in」，只換機位高度（地面、低角度、航拍），令剪接有一致節奏。
官方範例提示（英文原文）
Cinematic sequence of three shots with hard cuts, same glacial valley location, each shot is a slow steady push-in. Shot 1: ground-level wide shot from the valley floor, camera pushes in toward the glowing glacier at the far end of the valley, grass rippling in the foreground, mist drifting over the cliffs. Hard cut. Shot 2: low angle close to the dark basalt cliff wall on the right side, camera pushes in along the towering wet rock face with thin waterfalls streaming down, revealing the green valley opening below. Hard cut. Shot 3: elevated aerial angle above the valley floor, camera pushes in forward over the winding streams and vivid green grass toward the mist-covered glacier and the pool of light at center. Consistent look across all shots: photorealistic, cold cinematic color grade, bright diffused daylight, smooth stabilized dolly movement, no camera shake, no people, no animals, no birds.
影片模型 Seedance 2.5｜黑色電影：陰影與氛圍（五段鎖定鏡頭）
黑白、硬主光、每段鎖定鏡頭。年份＋城市＋天氣一句定下時代感。
官方範例提示（英文原文）
SCENE CONTEXT: Night, 1947, a rain-soaked downtown street in an American city. A trench-coated detective waits under a streetlamp outside a small hotel, a dark sedan pulls up, a woman in 1940s evening wear steps out, walks to him, delivers one line, and the two walk together through the rain toward the hotel doorway. LOCATION MAP: Camera works from the south sidewalk, facing north, for the entire sequence. Background, north side: dark brick hotel facade, a vertical neon sign reading HOTEL glows through rain haze on the right half of frame, a recessed doorway spills bright light onto the wet pavement below the sign. Midground: cast-iron streetlamp on the north sidewalk, 5 meters screen-left of the hotel doorway, a steam grate breathes near the kerb. Foreground: wet asphalt with standing puddles reflecting the neon and the lamp. Rain falls steadily with a slight leftward drift, atmospheric haze deepens with distance. FIRST FRAME AND SPATIAL BLOCKING: The first visible frame already contains THE DETECTIVE under the streetlamp with the full street geography readable. No empty establishing frame, no delayed reveal. THE DETECTIVE: man around 40, weathered face, day-old stubble, rain-darkened double-breasted gabardine trench coat belted at the waist, wide-brim felt fedora with grosgrain band, wide-lapel 1940s suit, cuffed trousers, leather oxfords, unfiltered cigarette in his right hand. Midground left third, x 35%, y 55%, body facing camera, hat brim shadowing his eyes. THE WOMAN (first appears at 12.0s): around 30, dark 1940s wave-set hair pinned under a small tilt hat with a birdcage net veil, dark lipstick, belted wool coat with padded shoulders over an evening dress, leather gloves, seamed stockings, ankle-strap heels, small clutch in one gloved hand. THE SEDAN (first appears at 7.0s): rounded-fender late-1940s four-door sedan, dark paint, chrome bumpers, round headlights, no badges, no emblems, no lettering. FORMAT MODE: Controlled five-segment multi-shot sequence, 30.0 seconds total, four HARD CUTs, real-time motion throughout. Screen direction never flips, the sedan arrives from screen-right, the woman travels right-to-left. No subtitles, no captions, no on-screen text other than the neon word HOTEL. OPTICS: Ultra-realistic black-and-white large-format capture, IMAX-scale negative clarity through Panavision-style anamorphic glass. LENS LOCK SEGMENT 1: 84° diagonal field of view, classic wide, camera about 6 meters from the detective at the start of the push. LENS LOCK SEGMENT 2: 29° diagonal field of view, short telephoto portrait, camera 4 to 6 meters. LENS LOCK SEGMENT 3: 47° diagonal field of view, standard normal, camera 3 to 5 meters, natural human-eye perspective. LENS LOCK SEGMENT 4: 18° diagonal field of view, classic telephoto, camera 6 to 8 meters, strong compression stacks the two faces close together. LENS LOCK SEGMENT 5: 84° diagonal field of view, classic wide, camera low near the ground. No lens drift mid-segment. CAMERA: Heavy studio dolly and locked-off tripod behavior, slow, deliberate, machine-smooth moves, no handheld jitter, no drone float. Segment 1: chest-height dolly, slow straight push-in from 6 meters to 3.5 meters. Segment 2: locked-off at chest height, focus holds on the detective while the sedan enters soft behind. Segment 3: chest-height dolly tracking screen-left with the woman. Segment 4: locked-off telephoto two-shot, both profiles in the same focus plane. Segment 5: static locked-off low angle, lens 40 centimeters above the wet asphalt. ACTION TIMING: 0.0s to 6.0s, SEGMENT 1, 84° wide. THE DETECTIVE motionless under the lamp, neon HOTEL sign glowing, steam drifting from the grate. At 3.5s he raises the cigarette and draws, the ember flares bright, he exhales and smoke curls through the lamplight. HARD CUT. 6.0s to 12.0s, SEGMENT 2, 29° medium. At 7.0s twin headlight beams rake across him. At 9.0s to 11.0s the sedan rolls in and settles at the kerb, the suspension dips and recovers. HARD CUT. 12.0s to 19.0s, SEGMENT 3, 47° normal. The rear door swings open, THE WOMAN rises out of the car, straightens her coat, pushes the door shut. She walks right-to-left toward the lamp, covering 4 meters, heels clicking a steady rhythm. HARD CUT. 19.0s to 25.0s, SEGMENT 4, 18° telephoto two-shot. At 20.0s THE WOMAN says: "You're late." Her voice is low, dry, and controlled. At 23.0s he drops the cigarette, it bounces once and dies with a brief hiss in a puddle. HARD CUT. 25.0s to 30.0s, SEGMENT 5, 84° low wide. They turn together and walk away from camera toward the hotel doorway, side by side, shrinking into the rain haze. They reach the spill of doorway light, the neon flickers once. PHYSICS: Rain falls at constant density and direction for all 30 seconds, drops strike puddles with rings and micro-splashes. Walking carries true weight, heel strike, weight transfer, hip shift, toe push-off. The sedan has real mass, it decelerates progressively, the body dips on its springs at the stop. The cigarette ember brightens with each draw and dies instantly in water. LIGHTING: Black-and-white low-key noir. Primary key: the streetlamp directly above THE DETECTIVE, a hard narrow overhead cone that carves the hat-brim shadow across his eyes until he lifts his chin at 8.5s. Secondary source: the neon HOTEL sign. Event source: the sedan headlights raking screen-right to screen-left from 7.0s to 11.0s. Exposure priority: expose for the highlights and let faces fall into hard shadow relieved only by eye glints and rim light. The entire image is monochrome silver-gelatin black and white from first frame to last, with fine period film grain and zero color anywhere. AUDIO: Continuous rain on pavement, awnings, and the sedan roof, gutter runoff, one distant thunder roll at 5.0s, neon buzz near the sign, steam hiss near the grate. Her heels click in steady rhythm from 15.0s to 19.0s. A faint muffled solo trumpet drifts from inside the hotel doorway, distant and diegetic, no score, no added music, no narration. The only spoken words in the sequence are THE WOMAN's line at 20.0s: "You're late." No subtitles. POSITIVE LOCKS: Both characters keep the same face, wardrobe, and rain-wet state across every cut, wetness only accumulates, never resets. Every visible element is period-correct for 1947, nothing modern appears in any frame. The sedan carries no brand marks, the only readable text anywhere is the neon word HOTEL. Left-right geography holds through all cuts, the woman always travels right-to-left, and the camera never crosses to the north side of the street.
影片模型 Seedance 2.5｜多角色：跨鏡保持多張臉一致（K-pop 四人組）
官方結論：多角色最難保持一致——人臉越多，每張臉需要的參考素材越多，否則第三、四刀後會漂成「通用臉」。平均鏡長 < 1.5 秒、踩拍硬切。
官方範例提示（英文原文）
SCENE CONTEXT: A four-member East Asian girl group in a super-fast hard K-pop hip-hop piece, told as one linear story in four acts: ORIGIN (a retro TV, a plunge into the screen, a retro-futuristic cockpit, space, a surreal dentist pocket reached through an eye), ARRIVAL (descent to Earth, a night airfield landing, exit in spacesuits), THE GAUNTLET (a huge crowd of fans and reporters they push through while performing), and THE DANCE (the beat drops as they break free and perform a group dance high on a tower, filmed both cinematically and by a circling reporter news helicopter). The clip lives in polished cinematic quality except during the gauntlet and the helicopter aerials, which are degraded reporter-camera footage. No cyclorama, no neon anywhere, no national flags. MEMBER COUNT LOCK (critical): The group has exactly four members, never five, never six. No fifth member, no extra unnamed dancer, no background double, no duplicated or cloned member. Each of the four appears once and only once in any frame, no mirror, glass, or reflection showing an extra copy. DUAL-QUALITY SYSTEM (critical): [FILM] is polished modern digital cinema, clean, sharp, stable, glossy, graded, deep glossy blacks, the default for the whole clip. [REPORTER CAM] is a degraded early-2000s cheap phone, camcorder, or news-helicopter look, heavy VHS-grade compression, chunky macroblock artifacts, smeared low resolution, muddy color, blown-out highlights, constant handheld or chopper jitter, crude digital crop-zoom punches, autofocus hunting, rolling-shutter wobble. The degradation appears only in REPORTER CAM beats and never leaks into FILM beats. TRANSITION MOTIF: Alongside hard cuts, creative MATCH CUTs on round shapes and through eyes, a round CRT screen, a round porthole, a round dentist lamp, a planet disc, and a member's eye all rhyme. The camera pushes into the round shape or into an eye until it fills frame, then emerges in the next scene. FORMAT MODE: Controlled twenty-four-beat rapid sequence, real-time motion, hard cuts and a few match cuts on the beat, average beat around 1.25 seconds. Every beat has a moving camera and hard motion, no static holds. Each beat tagged FILM or REPORTER CAM. ACTION TIMING, approximately 152 BPM hard trap-drill K-pop, cuts land on the beat. ACT 1, ORIGIN: 0.0s to 1.25s, FILM, an old wood-panel CRT glows in a dark room, camera pushes into the screen until white fills frame, match cut into round screen. 1.25s to 2.5s, FILM, cockpit reveal, the four in white spacesuits hitting a unison pose. 7.5s to 8.75s, FILM, one member floats in zero gravity in full body against the star-field, the camera orbiting her floating form, ending pushed toward her eye, match cut into her pupil. 8.75s to 10.0s, FILM, out of the pupil into a bright clinical dental room, one member alone in the dentist chair, match cut through the round lamp. ACT 2, ARRIVAL: 10.0s to 11.25s, FILM, the ship punches down through cloud, hot orange heat glow raking the hull. 12.5s to 13.75s, FILM, the gear strikes the tarmac, steam and smoke billowing. ACT 3, THE GAUNTLET: 13.75s to 15.0s, REPORTER CAM, degraded phone POV from behind the crowd barrier, violent handheld shake, other phones jutting into frame. 15.0s to 16.25s, FILM, the ramp lowers, the four appearing backlit in the doorway in spacesuits. 17.5s to 18.75s, FILM, the lead member walks the four across the floodlit tarmac straight toward camera, exactly four in a line. 20.0s to 21.25s, FILM, camera body-locked to the lead member as she pushes forward through the packed crowd. ACT 4, THE DANCE (tower, card outfits, no spacesuits): 21.25s to 22.5s, FILM, all four burst into clear floodlit space and hit a hard unison hit as the beat drops, ending on a hard close of one member's eye, match cut into eye. 22.5s to 23.75s, FILM, out of the eye onto a tall open steel platform atop a broadcast tower high above the night city, the four now in their card outfits dance hard in a line, exactly four, no double. 23.75s to 25.0s, REPORTER CAM, degraded news-helicopter footage, the four small on the distant tower platform dancing. 28.75s to 30.0s, FILM, final freeze, the four land the final unison pose, camera cranes up and away for the last frame, exactly four figures. PHYSICS: Real ground contact and weight transfer on every step, true jaw and lip mechanics on the rap, hair and cloth lag on spins and whips. Zero gravity in space, hair and necklace drift weightless with slow inertia. The ship carries real mass on descent and landing, gear compresses, steam billows. The crowd has real bodies and mass, fans and reporters press, sway and reach with true weight. FILM camera moves carry real inertia, REPORTER CAM motion is human or chopper chaos, real hand shake, stumble, jostle, never gimbal but plausible. Exactly four members and five fingers per hand. LIGHTING: FILM beats lit for grounded cinematic contrast, never murky, no neon anywhere. TV room near-black with CRT glow only. Cockpit lit in cool cyan and warm amber. Space is black star-field with cool blue-white rim light. Dentist bright clinical white with one cold-blue accent. Tower platform cool moonlight plus hard tower floodlights plus warm distant city glow. REPORTER CAM beats show the same physical lights through a cheap sensor, crushed muddy shadows, harshly blown highlights. AUDIO: Music added in post, no generated vocals, no words, no melody from the model. Mouth motion is rap cadence, not synced to specific words. Diegetic bed only: CRT hum, ship thruster rumble, steam hiss, crowd roar, helicopter rotor thrum under the aerials. POSITIVE LOCKS: The group is exactly four members across all beats, no fifth member, no duplicate or cloned member. Faces and hair stay fully consistent with the four reference images. Wardrobe by act: spacesuits in origin, arrival, and gauntlet, card outfits on the tower. Spacesuits carry no national flags, no flag patches. No neon anywhere. The dual-quality system is strict, FILM beats stay fully clean, REPORTER CAM beats carry the full degraded look, the two never mix in one shot. Real-time 24fps, no slow motion, no speed ramps, no ghosting or trails.
影片模型 Seedance 2.5｜奇幻動作：角色與巨獸的尺度
「Hyperbolic speed, no blood anywhere」——把排除項（無血）寫進設定，廣角＋長焦混用表現尺度。
官方範例提示（英文原文）
SCENE CONTEXT: A woman knight in silver armor dives on a winged white horse from a burning tear in the sky down into a black volcanic canyon. One flying demon tries to stop her and she takes its head off. She levels out over a rock ledge crowded with demons, rides through them, tries to cut the giant archdemon and her sword skates off his armor. When he swings his huge blade and misses, his side opens up, and she slices it open as she flies past. CHARACTER: A woman in her mid-twenties, a warrior, focused and calm, not afraid. Silver etched plate armor over a teal-blue quilted underlayer, teal-blue quilted skirt panels hanging to mid-calf, brown leather belt, two tight braids running back along her scalp into long black braids, pale blue-grey eyes, silver ear cuff, a straight double-edged longsword with a plain steel crossguard and dark leather grip. No helmet, no cloak. Everything about her, face, skin, hair, armor, sword, comes from the reference image and must stay the same in every shot. Through the sequence she gets dirtier and never gets clean again: soot on her face from the start, then grey ash and glowing ember dust caught in her braids and armor after the first cut, a scorch mark and burnt streaks across her chest and left arm, grey dust down her left side after she rides through the horde, a nick in her sword blade after it hits the archdemon's armor. No blood on her, no blood on anything. HOW DEMONS DIE, NO BLOOD ANYWHERE: There is no blood in this film, not one drop, from any creature. Demons are made of cooled black lava with fire trapped inside them. When the sword opens one up, the cut edge glows bright orange for a moment as the fire inside escapes. A burst of sparks and hot embers blows out of the wound, along with a jet of shimmering heat. Then the body loses its light, goes grey, cracks apart, and crumbles into ash that the wind instantly tears away. Never liquid, never red, no spraying, no pooling, no gore. HOW SHE USES THE SWORD, MOST IMPORTANT RULE: She only cuts. She never stabs. There is not one stab, thrust, lunge, or point-first movement anywhere in these thirty seconds. The sharp edge always goes first, the point always drags behind. Every cut has four parts: she pulls the sword back and her shoulder loads up, she turns her upper body and the sword sweeps forward in a curve edge first, the edge enters and keeps moving along and through the target, her arm keeps going past the target and finishes fully extended. The sword never stops inside a body. There are only five sword actions in this film: at 6.4s she draws it, at 9.2s she takes the flying demon's head off, at 15.6s she cuts through two demons on the ledge, at 17.2s she tries to cut the archdemon's shoulder and it fails, at 24.6s she slices open his side. SLOW MOTION, ONLY TWO MOMENTS: The whole film runs at real speed except two short beats, both about her blade meeting the archdemon's body. Slow beat one, 17.2s to 17.7s, her sword hits his shoulder armor and skates off in a shower of sparks. Slow beat two, 24.6s to 25.6s, the edge enters his bare side and draws along it. Nowhere else, no other slow motion, no speed ramps. SPEED: The horse flies fast, around 65 meters per second, like a falcon diving. Its wings stay swept back and half folded most of the time, opening only five times in the whole film. Her braids, the horse's mane and tail, and her quilted skirt panels are all pressed flat backwards by the wind. Shot at 24fps with normal motion blur. THE OTHER CHARACTERS: THE HORSE, one white winged stallion, big and heavy, about 700 kg, off-white coat with grey dapples, feathered wings with a 6-meter span. THE ARCHDEMON, one giant demon, 5 meters tall, three times as tall as any other demon, black volcanic rock fused onto muscle, cracked across the chest with slow orange fire glowing underneath, eight black horns, a huge two-handed cleaver of black volcanic glass 2.6 meters long. He is slow and heavy, every move takes a long wind-up. HIS WEAK SPOT: his rock armor stops at his ribs and starts again at his hips, a strip of bare cracked skin about 90 cm long, glowing dull orange, normally covered by his hanging arm and cloak, only opening when he swings hard and his body twists. THE FLYING DEMON, one only, 2.2 meters tall, dies at 9.2s. THE HORDE, hundreds of smaller demons on the ground, 2.2 to 2.5 meters tall. THE PLACE: A huge volcanic canyon, kilometers deep. Above, a torn burning gap in black storm clouds with blinding gold-white fire pouring through. Below, enormous black rock spires, 40 to 90 meters tall. Colors: almost black and white, cold grey-blue everywhere, with gold fire as the only real color. DIRECTION: The fire gap is always in the upper left. She always flies from left to right and downward. The archdemon stands at the lower right, facing left toward her. The camera always stays on the same side, her right side, his left side. HEIGHTS: The ledge floor is zero. The horde demons are 2.5 meters tall. The archdemon is 5 meters tall, his eyes at 4.5 meters. She levels out at 4 meters above the ledge at 12.8s, drops to 3 meters at 21.5s, below his eyes so during the final pass he is looking down at her. FORMAT: Seventeen shots, 30 seconds total, sixteen cuts, averaging under two seconds each. Every cut lands on movement, never on a still pose. CAMERA AND LENSES: Anamorphic look throughout, oval out-of-focus highlights, slight softening at the frame corners. In every shot the camera is struggling to keep up with her, it lags behind, over-corrects, catches up again. Handheld weight, operator breathing, real human correction, no smooth drone glides. SHOT 1, THE CANYON, 0s to 2.2s, wide: Camera low on a foreground rock, tilted up toward the burning gap. She is a small pale dot high in the upper left, crossing in front of the fire, dropping toward the lower right. At 1.4s the horse gives one hard downbeat. SHOT 4, THE DRAW, 5.6s to 7.2s, medium: At 6.4s she draws the sword, one single sweeping movement, her right hand grabs the grip at her left hip, pulls the sword up and across her body in one continuous curve, edge leading the whole way, ending held down and back behind her right hip. SHOT 17, final shot: The last cut lands on the finished slice through the archdemon's side, sparks and ash blowing away in the wind as she flies past. PHYSICS: Real weight and inertia throughout. The horse has genuine mass in flight, wingbeats push real air. Her sword swings carry follow-through and momentum. Ash and embers rip past the camera much faster than the background moves. LIGHTING: Almost black and white, cold grey-blue everywhere, with gold fire as the only real color. Long straight beams of light fan down through the haze at a steep angle from the burning gap above. Cracks of dull red lava between the rock spires. AUDIO: not specified in the source, follows the same real-time, no-music convention as the sequence's visual restraint.
影片模型 Seedance 2.5｜UGC 廣告：自然動作、無製作感（30 秒直向）
指定「shot on iPhone 14 Pro, authentic phone-camera look」與固定自拍構圖，令畫面像真人手機拍攝。
官方範例提示（英文原文）
Vertical 9:16 UGC video, 30 seconds total, 24fps, shot on iPhone 14 Pro, authentic phone-camera look, handheld selfie framing from her seat for the whole video, arm's-length wobble, tiny refocus moments, soft even cabin lighting, natural vlog pacing, 8 shots, one consistent framing with small variations, the first shots shake with real turbulence, then the frame settles as the flight smooths out. Cast: the young woman from the reference image, face fully visible and matching exactly in every shot, fair skin, shoulder-length pastel-pink wavy hair under a burgundy beanie with a small silver star pin, blue eyes, small stud earrings, natural minimal makeup. Her acting is understated and real, she blinks at a natural rate, reactions arrive a half-beat late, smiles start small and grow, she glances away from the lens and back mid-sentence, and she keeps her voice down like someone slightly self-conscious about filming on a plane, no wide eyes, no exaggerated faces, no performing. Background passengers, realistic, alive, never NPC-like: scattered through the blurred cabin, a middle-aged man asleep against a window with his head tilted and mouth slightly open, a woman in her 30s reading a paperback, an older man scrolling a phone with reading glasses low on his nose, a young woman with headphones gazing out the window. Each moves subtly and independently, nobody looks at her or her camera, nobody reacts in sync. Product: the exact box from the reference, BAKEY Chocolate Chip Cookie Dough, The London Bakehouse Co, Ready to Bake, red box with cream lettering and the cookie photo, sitting on her tray table. Location: the airplane cabin from the reference, navy leather seats with white GGS AIR headrest covers, grey armrests, oval windows with soft daylight. Seating geometry fixed for every shot: she sits in the aisle seat, the aisle at her left shoulder, the seats to her right toward the window occupied by the sleeping man. Segment 1 (0-10s): Shot 1, Turbulence Hook (0-4s), handheld selfie, the frame shaking, the cabin rattles through rough turbulence, she grips the armrest with one hand, phone wobbling in the other, and talks to the lens through it in a tight controlled voice: "So we are currently being shaken like a snow globe." Shot 2, It Settles (4-7s), same framing, the shake easing beat by beat until calm, she exhales slowly and says: "Okay. We're fine. Everything's fine." Shot 3, The Pivot (7-10s), she looks down at her tray table, lifts the red box into frame, pats it twice, and says: "And this survived, which is all that matters." Segment 2 (10-20s): Shot 4, The Open (10-13.5s), closer selfie, she folds the box open and brings out one golden cookie, no speech. Shot 5, First Bite (13.5-17s), she takes an unhurried bite, chews for a real moment, and says: "Okay, that's actually really good." Shot 6, The Detail (17-20s), closer on the cookie, she breaks it in half slowly and murmurs: "Look at those chocolate chunks." Segment 3 (20-30s): Shot 7, The Verdict (20-25s), same selfie framing, relaxed, she gives her recommendation: "These are Bakey, by the way. Honestly the best thing I packed for this trip." Shot 8, Quiet Outro (25-30s), no more words, she finishes the half in two calm bites, dusts her fingertips, tucks the box flap closed, and lets her head rest back against the seat, the frame holds calm and still, no product packshot. Material realism: the turbulence reads true, the handheld frame jolts in irregular bumps, hair and tie sway with the plane's motion. The calm afterward is equally true, the frame's motion decays gradually, not instantly. The cookies are golden with matte cracked surfaces, soft flex at the bite, dark chocolate chunks glossy, real crumbs on the napkin. Skin is real human skin, visible pores, natural sheen, no plastic smoothness. Audio: no music, she speaks on camera with accurate lip-sync. The soundscape carries the story, rough rattling cabin, creaking bins and a seatbelt chime under Shot 1, the rattle decaying into smooth engine hum in Shot 2, the cardboard flick in Shot 4, the soft cookie snap in Shot 6. Her voice: young American accent, warm and low-key, relaxed unhurried pace, every line a complete finished sentence, no trailing off. All speech ends by 25s. Consistency rules: her face stays the same person from the reference in every frame. She stays in the same aisle seat for the entire video, no seat changes, no teleporting. Turbulence exists only in Shots 1-2 and never returns. The background passengers keep their same seats, faces, clothes and activities in every shot. The BAKEY box lettering stays sharp and correctly spelled whenever visible. The cookie only shrinks, whole in Shot 4, bitten from Shot 5, two halves in Shot 6, finished in Shot 8, never regrowing. Exactly five fingers per hand. No on-screen text or captions, no product packshot at the end, the video ends on her.
影片模型 Seedance 2.5｜恐怖：錯的光、錯的角度（4:3 一點透視）
以 Physics 段開頭，強調真實重量與聲音（波鞋在膠地上吱吱聲、貨架上的穀物盒倒下），用 4:3 與深一點透視製造不安。
官方範例提示（英文原文）
Physics: real weight and inertia, sneakers squeak and slip on linoleum, cereal boxes topple with real mass, bag contents shift naturally, the stockroom door slams with impact. Composition: deep one-point-perspective aisles, the girl small against long empty corridors, the negative space past her shoulder held empty, then filled. Continuity: the girl, her outfit and bag identical across all shots and matching the reference, the supermarket matching the reference, the masked antagonist identical across the film and matching the reference. Technical: 24fps, 4:3 aspect ratio, total runtime 30 seconds. SUBJECT: an 18-year-old East Asian girl with a soft round face, no makeup, tired eyes, messy dark brown hair pulled into a loose bun with wispy bangs and loose strands falling across her face, wearing a plain faded gray oversized hoodie with no logos or patches, loose blue jeans, worn cream canvas sneakers, a small dark canvas tote bag slung across her shoulder. Fully alive human micro-acting throughout: real blinking, chest heaving, throat swallows, tears welling without falling. ANTAGONIST: a tall, very thin young woman with pale human skin wearing a full-face expressionless matte white doll-like mask with dark hollow eye openings and a small painted dark rosebud mouth, messy dark brown shoulder-length hair with choppy bangs, two red fabric flowers pinned in her hair, a dingy knee-length vintage cream slip dress, thin bare arms, long thin fingers, bare pale feet. Rule: she is never in sharp focus until the final second, always a soft blur, a silhouette, a fragment behind shelves, the autofocus refuses to lock on her. The mask never comes off.
影片模型 Seedance 2.5｜紀錄片：單一 47° 固定鏡頭，只改距離
全片同一 47°（接近人眼）視角，只改攝影師站位距離；以「POSITIVE LOCKS」段落列出必須不變的事項（海在右、沙丘在左、同一女子的臉／雀斑／疤痕／泳衣）。
官方範例提示（英文原文）
SCENE CONTEXT: A large university beach party on the same stretch of coast, followed from a sunny afternoon through the night and into the blue hour before sunrise. We follow one young woman through five moments, she runs into the surf in bright daylight, dries off and laughs with friends on the sand, dances in the crowd at the bonfire after dark, stands alone at the waterline late in the night when the party has thinned, and watches the sky go pale blue at dawn. Nothing dramatic happens, it is one long good night, observed by someone standing on the beach with her. FORMAT MODE: Controlled multi-shot sequence, 30 seconds total, 2.39:1 widescreen, third-person observational, live-action photoreal. Five segments, four hard cuts at fixed times. Real-time motion, no slow motion. No subtitles, no on-screen text. Diegetic sound only. Ends on the final frame with no fade. TIME AND CROWD PROGRESSION LOCK: This is one unbroken span of time on one beach, not five different days. The party swells and then empties, the crowd size is the clock. Segment 1, bright mid-afternoon, roughly forty people. Segment 2, late afternoon, roughly sixty, fires being built. Segment 3, full night, peak of the party, roughly a hundred people packed around the main bonfire. Segment 4, deep night hours later, fifteen or twenty left. Segment 5, blue hour before sunrise, five or six remain. The same dune line, the same two fire pits, and the same accumulating debris persist across all five segments. CHARACTER: A 23-year-old university student, slim athletic build, roughly 172 cm. Identity anchors identical in every segment: long dark blonde hair, thick and naturally wavy, worn in a loose knot that keeps falling apart. High cheekbones, a straight nose, clear grey-blue eyes, a scattering of freckles across the nose bridge, a small pale scar through the left eyebrow, a thin braided cord bracelet on the right wrist, permanently on. BEAUTY WITHOUT RETOUCHING LOCK: her skin is fully unretouched and shows everything, visible pores, a developing sunburn, salt drying to a faint white bloom, sand stuck along the forearms, tired puffy eyes by dawn. No airbrushing, no skin smoothing, no glamour retouching, no beauty filter. WARDROBE, layered progressively: Segments 1 and 2, plain terracotta one-piece swimsuit, a loose faded white linen shirt worn open over it. Segment 3, the same plus a heavy oversized oatmeal knitted cardigan. Segment 4, the same, cardigan wrapped tight. Segment 5, the same plus a striped wool blanket around her shoulders. The terracotta swimsuit and the dark blonde knot are the visual anchor that identifies her at any distance. FRAMING RESPECT LOCK: The camera treats her as a person at a party, not as a body to be looked at. It stays on her face and her actions. No lingering framing of the body, no slow pans up or down a figure, no isolated shots of torso, hips or legs, no low camera angles. IDENTITY LOCK: Every person in this sequence is fully fictional. No face resembles any real, living or historical person. All are university-age adults, early to mid twenties. NO-IP LOCK: No logos, no trademarks, no brand names, no readable signage anywhere in frame. The party music is original, generic and non-descript, no vocals, no recognisable melody. Nobody speaks an intelligible line. CLEAN-AIR LOCK: Nobody in this sequence smokes, vapes or exhales anything visible. The only smoke in the entire sequence comes from the two fires, and it always travels out to sea, away from the people. BEACH MAP: The camera stays on the landward side of the action for the entire sequence, the sea is always screen-right, the dune line always screen-left. This never reverses. SINGLE-LENS LOCK: One focal length for the entire 30 seconds, 47° diagonal field of view, standard normal lens character. This never changes, in any segment. Every change in framing is achieved by the operator physically walking closer to or further from her. CAMERA-TO-SUBJECT DISTANCE PER SEGMENT: Segment 1, 5 meters. Segment 2, 2 meters. Segment 3, 3 meters. Segment 4, 2.5 meters. Segment 5, 7 meters. ACTION TIMING: 0.0s to 7.0s, Segment 1, into the surf, camera 5 meters. She is running down the wet sand toward the water, hits the shore break at 2.0s, laughing with her whole face. Hard cut. 7.0s to 13.0s, Segment 2, drying off, camera 2 meters, operator sitting on the sand. She re-ties her hair, takes a drink from an unmarked bottle. Hard cut. 13.0s to 20.0s, Segment 3, the fire, camera 3 meters, standing inside the crowd. She is dancing loosely in the middle of the crowd, firelight flickers across her face. At 16.0s a log collapses and sparks lift. Hard cut. 20.0s to 26.0s, Segment 4, the waterline, camera 2.5 meters. She stands ankle-deep at the tide line, looking out at the black water, stands still for three full seconds. Hard cut. 26.0s to 30.0s, Segment 5, blue hour, camera 7 meters, operator crouched. She sits on the cold sand facing the sea, the fire a low mound of grey ash. Cut on the last frame, no fade. PHYSICS: Water behaves with real mass, the shore break stops a running body hard, spray leaves the feet in flat sheets. Wet fabric clings, darkens, and hangs heavier. Soft dry sand gives under every step. The fire has real heat behaviour, convection lifting sparks in an unsteady column. LIGHTING, five states, one coast: Segment 1, high afternoon sun, hard and almost overhead. Segment 2, sun dropping, light warming and raking in low. Segment 3, night, the bonfire is the dominant source, low, warm, orange, unsteady. Segment 4, deep night, primary source is a low moon, laying a soft silver path across the water. Segment 5, blue hour, no sun above the horizon, a single vast soft even source, the whole sky. AUDIO: 100 percent real on-location sound, no score. The surf is the continuous through-line across all five segments. Music is diegetic only, from the speaker stack, original generic instrumental, no vocals. No intelligible dialogue. POSITIVE LOCKS: Live-action photoreal only, real unretouched skin. The 47° natural human-eye perspective is identical in all five segments, only the operator's standing distance changes. The sea stays screen-right and the dune line screen-left in every segment. The same woman is present in every frame, her face, hair colour, freckles, eyebrow scar, and terracotta swimsuit stay identical across all five segments.
影片模型 Seedance 2.5｜CLI：t2v 與 omni_reference
官方 README 範例（譯：日出時山谷上空的航拍鏡頭）。圖生影改用 --mode omni_reference --start-image ./first.png（Skills 範例：--prompt "camera dollies in" --duration 12 --resolution 1080p）。
官方範例提示（英文原文）
higgsfield generate create seedance_2_5 \ --prompt "drone shot over a mountain valley at sunrise" \ --aspect_ratio 16:9 --duration 5 \ --resolution 1080p --mode t2v --bitrate_mode high \ --wait
影片模型 Seedance 2.0（含 2.0 Mini）｜Pro tip：怪物／寫實質感太光滑時
官方：怪物或任何要「感覺真實」的東西，皮膚太滑太塑膠時，在提示加上這句。
官方範例提示（英文原文）
no 3D, no cartoon, no VFX
影片模型 Seedance 2.0（含 2.0 Mini）｜Pro tip：喜劇背景笑點
官方：寫這句，Seedance 會自行發明一個背景笑點，不用你描述。
官方範例提示（英文原文）
add a visual gag in the background
影片模型 Seedance 2.0（含 2.0 Mini）｜Pro tip：特效寫在動作描述內的方括號
官方：把特效以方括號內嵌在動作描述中，告訴模型特效出現的確切時機與位置。
官方範例提示（英文原文）
[VFX: branching electric circuits pulsing with white-blue current]
影片模型 Seedance 2.0（含 2.0 Mini）｜Pro tip：POV 要寫明鏡頭「不做」甚麼
官方：單憑這句就能把第一身視角鎖住。
官方範例提示（英文原文）
No cuts, no zoom, natural head movement
影片模型 Seedance 2.0（含 2.0 Mini）｜變身格式（6 鏡 15 秒，4 張參考圖）
官方範例「The Burger」：吃漢堡的女孩變成巨獸吞掉殭屍再變回來繼續吃。結構：總體質感行 → 場景段 → Shot 1–6。輸入：Image × 4（女孩、貨車、地點、殭屍）。
官方範例提示（英文原文）
Montage, multi-shot action Hollywood movie, don't use one camera angle or single cut, cinematic lighting, photorealistic, 35mm film quality, professional color grading, sharp focus, high detail texture, film grain, depth of field mastery, ARRI ALEXA aesthetic A pink-haired girl with glasses, cream top and jeans sits on the hood of a white pickup truck under a concrete overpass at dusk, casually eating a burger. A shallow river channel stretches behind her, power lines and distant bridges framing the golden sky. A pale zombie with wet dark hair, bruised eyes and a blood-stained white shirt sprints toward her from the shadows of the channel. The girl calmly sets down the burger, her body erupts into a massive pale tusked creature with elongated limbs and clawed hands, devours the zombie whole, then shrinks back to human form and picks up the burger. Handheld shake throughout, dark comedy pacing with horror undertones. Shot 1: Medium shot of the girl sitting cross-legged on the truck hood, chewing the burger lazily, golden dusk light catching her glasses and pink hair. Camera sways gently, ambient and calm. Shot 2: Wide shot of the concrete channel as the zombie bursts from the shadows under the bridge, sprinting with jerky unnatural strides across the dry riverbed toward the truck. Camera shakes tracking the approaching threat. Shot 3: Close-up on the girl's face as she notices the zombie, chewing slows, eyebrows rise with mild annoyance rather than fear. She sets the burger down on the hood beside her. Shot 4: Medium shot as the girl drops off the hood and her body violently expands and twists upward into the massive pale tusked creature, spine cracking, limbs stretching, jaws splitting open wide, towering over the truck. Camera jolts with each bone-snap of the transformation. Shot 5: Wide low-angle as the creature lunges forward and catches the charging zombie in its enormous clawed hand, lifts it off the ground and swallows it whole in one grotesque bite, jaw unhinging. Camera shudders with the impact. Shot 6: Medium shot as the creature rapidly shrinks back into the girl, standing calmly beside the truck. She hops back onto the hood, picks up the burger, takes another bite and keeps chewing as if nothing happened. Total: 15s / 6 shots / 16:9
影片模型 Seedance 2.0（含 2.0 Mini）｜能量球（Orb）格式：第一身 + 混亂手持
官方「Orb — Electro」：一鏡到底第一身（the camera IS her eyes），特效以 [VFX:] 內嵌。
官方範例提示（英文原文）
Single continuous shot, first-person POV perspective, the camera IS her eyes, hyper-chaotic handheld motion, completely unstabilized, violent raw human movement, constant micro-jitters, aggressive head swings, abrupt jerks, frequent over-rotation and harsh correction, moments of near motion blur loss, no smoothness at all, no stabilization, wide-angle lens (strong distortion), subtle chromatic aberration near frame edges, 15 seconds, her hands always visible in frame, no music only raw SFX, cinematic lighting, photorealistic, grounded realism, strong 35mm film look, heavy film grain, sharp but imperfect focus, noticeable focus breathing, motion blur on fast actions, halation on highlights, soft highlight rolloff, slightly desaturated tones, ARRI ALEXA aesthetic, practical VFX feel, minimal CGI look, natural imperfections In a storm-soaked industrial ruin, her hand snaps forward catching a crackling violet lightning sphere suspended between broken metal beams, electricity arcing violently between surfaces, she crushes it instantly — blinding energy surges through her fingers as fractal lightning veins explode across both forearms [VFX: branching electric circuits pulsing with white-blue current, sparks jumping between fingers], the ground trembles as enemies emerge — dozens of jagged obsidian creatures crawl rapidly across walls and ground, their glowing red cores flickering, while a colossal iron titan rises behind them, towering above ruined towers, its chest a massive rotating electromagnetic core crackling with energy, she lunges forward — first strike, she grabs a creature mid-leap and overloads it, its body bursting into a spray of glowing shards, second attack, she slams both hands down releasing a radial lightning surge that chains between multiple enemies, frying them mid-motion, the titan retaliates firing a massive beam of compressed energy, she sidesteps violently, catching the beam with one hand, redirecting it upward while sprinting straight at the titan, she runs up its collapsing body using magnetic pulls, reaches the core and drives both hands inside, unleashing a catastrophic surge, RAMPS TO SLOW MOTION as the core fractures, energy tearing outward in layered shockwaves, metal plates peeling away — SNAPS BACK, she rockets skyward, both lightning-glowing hands at frame edges, looking down as the titan collapses in a chain reaction explosion, electrical storms spreading across the battlefield below SFX: electric crackle, sphere hum surge, energy burst, crawling creature skitter, deep metallic titan rise, sharp discharge pop, lightning chain blast, slow-motion electric hum stretch, snap impact, beam charge roar, energy deflection crack, metal tearing, core overload rumble, slow-motion energy rupture, explosive collapse, upward blast whoosh, distant thunder roll, debris falling Total: 15s / 1 shot / 16:9
影片模型 Seedance 2.0（含 2.0 Mini）｜POV 格式：角鬥士
官方範例，示範「no cuts, no zoom, natural head movement」鎖定視角。
官方範例提示（英文原文）
One continuous shot, POV gladiator perspective in the Colosseum arena, no cuts, no zoom, natural head movement, a furious enemy warrior sprints straight toward the camera through thick dust and sunlight, heavy footsteps shaking the ground, as he reaches close the POV character reacts instantly — grabs him mid-charge and slams him violently onto the sand, powerful impact, dust explosion, no pause, immediate chaos erupts all around, multiple gladiators fighting simultaneously, intense close combat everywhere, swords clashing, shields smashing, bodies colliding, archers releasing arrows overhead, a chariot bursts through the scene, fighters falling and getting back up, constant aggressive motion, POV character fights barehanded with raw force — blocking strikes, grabbing opponents, throwing them aside, fast reactions, heavy hits, kinetic movement, camera shaking from impacts, breath sounds, dust in the air, dramatic sunlight beams cutting through shadows, epic scale, high intensity battlefield, cinematic, photorealistic, ultra detailed, motion blur on hits Total: 15s / 1 shot / 16:9
影片模型 Seedance 2.0（含 2.0 Mini）｜打鬥格式：火車頂一鏡 15 秒
官方範例：以「Single continuous shot 15s:」開頭，從極低角度起，逐段描述動作與鏡頭。
官方範例提示（英文原文）
Single continuous shot 15s: The camera begins below the train roof level — an extreme low angle looking up from beneath the rushing steel edge, the grey sky above and the train's metal roof surface filling the upper frame, power line cables visible on both sides, electric sparks already arcing between the cables and the roof surface in irregular white-blue bursts. The FPV arm accelerates upward and crests the rooftop edge just as both female samurai asian girls warriors sprint into frame from opposite ends of the carriage roof — dark armor catching the electric flash light, katanas already drawn and trailing behind them — the camera dropping to roof level and sweeping immediately into a full 360-degree orbit around both fighters as they clash center-roof, blades ringing against each other, feet sliding on the curved metal surface, the passing landscape blurring at speed on both sides and the electric sparks erupting in curtains around them with each power cable they pass beneath. The orbit completes and the camera settles tight behind the first warrior as she draws back her katana in a full overhead swing — the motion ramping into deep slow motion, every detail suspended: the blade arc, the electric spark frozen mid-burst beside her shoulder, the second warrior already reading the strike and arching her entire body backward in a deep spine bend, the katana blade sweeping through the space where her face was a fraction of a second before, the flat of the blade passing directly in front of her eyes at centimeter distance — and a single small lock of dark hair catching the blade edge and separating, the severed strand drifting upward in the slow motion wind. The motion snaps back to full speed as the second warrior completes her dodge and drives her elbow into the first warrior's chest simultaneously, the two fighters exchanging rapid blade-locked strikes across the carriage roof as the electric sparks shower around them. The camera pulls back to a wide rooftop angle as both warriors simultaneously lock arms, pivot, and drive each other outward over the edge — both bodies leaving the roof in the same moment, tumbling sideways off the carriage and falling through open air toward the river running parallel to the track below, the camera descending with them in a controlled drop, both fighters still gripping their katanas with arms extended toward each other during the two-second fall. Both bodies hit the river surface simultaneously in twin white explosions, and the camera plunges beneath the surface with them — the world going deep blue-green, sunlight filtering down in shafts from above, both warriors already moving through the underwater space toward each other, dark armor trailing bubbles, katanas extended, the fight continuing in the silent slow drift of the river current. NO MUSIC / NO SFX Total: 15s / 1 shot / 16:9
影片模型 Seedance 2.0（含 2.0 Mini）｜動畫格式：@image 作首幀與風格參考
官方範例：第一句就聲明「@image 是首幀與風格參考」，再定義動畫風格。
官方範例提示（英文原文）
@image is the first keyframe and style reference. Cinematic stylized 3D animation — photorealistic desert environment, stylized characters. Hero: young woman, white braided hair, blue sleeves and leggings, brown leather vest and belt pouch, light build — uses speed and agility. Monster: colossal cosmic entity — massive glowing spherical body radiating neon rainbow energy (green, pink, red, blue, orange), multiple long dark blue tentacles, small vicious face with green glowing eyes and wide mouth, surrounded by swirling dust storm. Setting: vast arid desert — red-brown rocky terrain, massive dust storm wall, dramatic split sky blue and storm grey. High FPS, realistic particle physics. 0–3s: WIDE SHOT from @image. Hero sprints directly at the monster — tiny against its colossal scale. Monster fires a tentacle down like a whip — hero slides under it, the impact craters the desert floor, shockwave of dust and debris. She bounces up, runs along the tentacle as it retracts, reaches the main body. Drives her fist into the glowing surface — rainbow energy repels her, blasts her backward 10 meters skidding across desert sand. She lands, digs in. Looks at her hand — it's glowing faintly. Realizes something. 3–6s: Hero charges again — this time ABSORBING the energy on contact instead of fighting it. Her blue sleeves begin pulsing with absorbed rainbow light. Monster fires two tentacles simultaneously — she grabs one with both hands, swings on it, uses momentum to launch herself at the second, grabs it mid-air. Now holding two tentacles — she pulls them together, ties them in a knot using her whole body as leverage. Monster confused — tentacles tangled. She runs up the knotted tentacles toward the face. Half-second slo-mo — her mid-air between tentacles, rainbow light streaming off her arms, the monster's glowing face filling the sky behind her — SMASH full speed. 6–9s: She reaches the monster's face — drives both glowing hands directly into its eyes. Massive rainbow energy discharge explodes outward — the monster screams, the sound sending visible ripples through the dust storm. It convulses — tentacles flailing wildly, smashing into the desert floor creating craters. She holds on as the face bucks and writhes. The absorbed energy in her arms surges — she releases it all at once directly into the monster's core. Explosion of pure white light from the impact point, rainbow colors fragmenting outward like shattered glass. 9–12s: The monster's rainbow energy begins flickering — unstable. It rears back, swelling larger — about to explode or unleash everything. Hero drops to desert floor, sprints away full speed. The monster's energy builds — visible pressure distorting the air around it, colors strobing. It RELEASES — massive omnidirectional energy shockwave tears across the desert, flattening everything, the dust storm wall blasting apart. Hero dives behind a rocky outcropping — the shockwave passes over, sand and light washing across the entire landscape. Silence. 12–15s: Hero emerges from behind the rock. The monster still floats — but smaller, dimmer, energy depleted. Its tentacles hang limp. It looks at her — green eyes flickering. She walks toward it slowly. Raises one glowing hand — still holding its energy. Holds it out toward the monster. The energy flows back — she returns it. The monster's colors restore softly, gently. It slowly descends toward the desert floor, tentacles settling. They face each other. A moment. Monster turns, drifts away across the desert into the remaining dust. She watches it go. Wind catches her white braid. Cinematic stylized 3D animation matching @image, photorealistic desert particle simulation, volumetric dust storm, rainbow neon energy VFX, absorbed energy glow on character, realistic sand physics, fast dynamic cuts. 2.35:1, 24fps. Total: 15s / 1 shot / 16:9
影片模型 Seedance 2.0（含 2.0 Mini）｜填空模板：角色場景
官方模板（2026-06）：角色以 @ 引用，最後以 SFX 寫環境聲。
模板
@character walks through [setting], [action], camera [move], [lighting], [mood]. SFX: [ambient sound description].
影片模型 Seedance 2.0（含 2.0 Mini）｜填空模板：產品鏡頭
官方模板：產品鏡以排除句「No people, no text, no logos.」收尾。
模板
@product sits on [surface] in [setting], [lighting], camera [move], [atmosphere]. No people, no text, no logos.
影片模型 Seedance 2.0（含 2.0 Mini）｜填空模板：建立鏡頭
官方模板：先寫地點／時間／天氣，再讓角色從指定方向入鏡。
模板
Wide establishing shot of [location], [time of day], [weather], camera [move], [atmosphere]. [Character] enters frame from [direction] and [action].
影片模型 Kling 3.0（含 3.0 Turbo）｜@character 元素＋時間碼分段（15 秒牛仔競技）
官方範例：首行定類型、長度、光線與 @character（並重述參考的關鍵外觀），之後以 0-1.5s、1.5-3s… 時間碼逐段寫動作，「HARD CUT」標明剪接，最後 Physics／Lighting／Style。
官方範例提示（英文原文）
Cinematic live-action rodeo film, 15 seconds, golden-hour sunlight, starring @character, the blonde cowgirl from the reference: long voluminous honey-blonde waves, cream felt cowboy hat, white crop top, baggy blue jeans with brown belt, brown embroidered western boots. 100% matches the reference in every frame. 0-1.5s: inside a wooden bucking chute, backlit by low sun, @character sits astride a massive black rodeo bull, right hand pressing her hat down; a handler's arm pulls the gate rope; dust hangs in the warm rim light. At 1.5s the gate swings open. 1.5-3s: side profile tracking shot: the bull lunges out of the chute into the arena, she rides one-handed with her hat hand raised, hair and fringed chaps whipping; sun flares through the dust, packed grandstand behind. 3-4s: HARD CUT to a tight backlit close-up of her face under the hat brim, calm confident eyes, golden edge light in flying hair; a whip-pan through white fabric transitions out. 4-8s: wide arena coverage shot from behind silhouetted spectators, their hats and raised hands framing the foreground: the bull bucks and spins in the center of a sun-hazed dirt arena, dust clouds blooming with each kick; she stays balanced, torso countering every buck, one hand high. Rodeo safety clowns in green shirts circle at a distance. 8-10s: frontal shot: the bull charges toward camera through the dust, she rides tall, hand on her hat, crowd roaring in the stands. 10-11.5s: she leaps off mid-buck, a dynamic tumbling dismount, dust exploding on landing. 11.5-13s: low ground-level shot from behind her legs: brown boots with spurs planted in the dirt in sharp foreground, the bull walks away into the haze screen-right, a clown waves it off screen-left. 13-15s: she runs toward camera laughing with motion blur, then stops in a medium hero frame: wide smile, one hand touching the hat brim, golden backlight, cheering grandstand behind, the exact reference face. Physics: real bull mass and gait, dirt kicked in arcs, hair and cloth follow momentum, dust lingers in the air. Lighting: low warm sun, strong backlight and lens flares, faces lifted by bounce off the dirt. Style: anamorphic western cinema, warm Kodak tones, natural film grain, handheld energy on wide shots, crowd ambience and hoof impacts, no subtitles, no on-screen text.
影片模型 Kling 3.0（含 3.0 Turbo）｜首幀控制：從首幀位置開始的連續一鏡
官方範例：CAMERA 段寫「starts exactly at the start-frame position」；LOOK 段要求與首幀顏色連續（粉紅恤衫與鋼灰色車保持首幀顏色）。
官方範例提示（英文原文）
0.0-1.5s: Following his glance, he lifts his arm off the wheel, swings his legs out and stands up beside the open door, squinting into the light. 1.5-3.0s: A man in a rumpled band uniform sprints up the driveway out of breath and PRESSES a battered brass trumpet into his hands, gesturing toward the lawn; a dozen more musicians hurry in with tubas, trombones and a bass drum, forming a ragged semicircle around the car, watching him expectantly. 3.0-6.0s: He looks at the trumpet, at them, then lifts it, sets the mouthpiece, cheeks filling, and BLOWS a huge bright opening note, shoulders rising with the breath, fingers dropping onto the valves. On his second phrase the whole band CRASHES in behind him, drum kicking, tubas thumping, the group stepping into rhythm. 6.0-8.0s: He walks two steps forward still playing, the band falling in around him, faces appearing at the mansion windows; the frame holds him mid-phrase, eyes closed, fully committed, as the shot ends. CAMERA: starts exactly at the start-frame position: outside the car, three-quarter front at chest height, framing him through the open door, loose handheld, drifting slightly right as the band assembles, then holding on him. No cut. PHYSICS: real trumpet technique: correct embouchure, visible breath support, valves moving with the notes; brass has real mass and swings on straps; the drum head flexes on each hit; the door swings naturally as he stands. LOOK: continuous with the start frame: bright clear daylight, balanced exposure, detail kept in the white facade, no blown-out highlights, neutral white balance, matte skin, the pink shirt and steel-gray car keep their exact start-frame colors, 35mm film grain, no CGI smoothness. No logos, no readable text. AUDIO: quiet neighborhood ambience, birdsong, hurried footsteps, one breathless shout, the single bright trumpet note, then the full brass band swelling in: live and slightly rough.
影片模型 Kling 3.0（含 3.0 Turbo）｜一鏡到底＋POSITIVE LOCKS（6 秒更衣室）
官方範例：以 POSITIVE LOCKS 列出必須保持的次序與限制（拉鏈不碰、眼神只對鏡頭一次、只有一個鏡頭運動、只有一人）。用正面「鎖定」代替負面詞。
官方範例提示（英文原文）
SCENE CONTEXT Six-second one-take: a red-haired girl in a navy tracksuit on the floor of a deep-red locker room: jacket-fix, room-scan, then a level stare into the lens as the handheld camera pushes in to a close portrait. ACTIVE REFERENCES @GIRL: 100% per reference: copper-red hair pulled back, loose strands; steel-blue eyes; navy track jacket, white chest panel, navy pants, dark-red sneakers. Seated, one knee up. Contained defiance: fix as armor-adjustment, scan as a fighter reading exits, the stare a statement. @LOC: per reference: deep-RED locker room: crimson panels, red lockers with chrome latches, white tile, one cool overhead pool into red gloom. FORMAT / CAMERA ONE UNBROKEN TAKE, 6.0s, zero cuts. First frame = reference composition, already alive: hands on the hem, eyes down, no empty start. 47 degrees natural, rectilinear, focus locked on her face. MAXIMUM HANDHELD: breath-sway, micro-corrections, framing by feet, never zoom, never stabilized. ONE continuous move: slow walking push 2.5m to 1.2m, waist-up to close portrait, at its closest exactly as her eyes land on the lens. ACTION TIMING 0-2.5s, THE FIX. Head down. Both hands tug the hem, squaring the white panel; collar flick, a copper strand falls across her temple; she leaves it. The zipper stays untouched, fingers never near it. Camera breathes, then advances on the collar flick. 2.5-4.5s, THE SCAN. Hands settle on the knee. Eyes sweep LEFT into the red depth, hold, track RIGHT along the chrome latches: reading the room, not the lens. Camera closing. 4.5-6.0s, THE LOOK. Eyes come off the lockers, land level on the lens. Chin lifts a degree. One slow blink, then stillness, breath in the shoulders. Camera settles into close portrait, holds. Ends MID-STARE. PHYSICS / LIGHT / AUDIO Real breath, true blink timing, nylon rustle, the strand staying. One cool overhead key carves cheekbones and the white panel; red gloom, red bounce, chrome speculars; her face takes more key as the camera nears. Ventilation hum, fabric rustle, one nose-exhale. No music, no dialogue. POSITIVE LOCKS Order locked: fix (zipper untouched), scan left-right, eyes land on the lens and HOLD. Eyes meet the lens exactly ONCE, at the end. ONE camera move. ONE person; 100% per reference. 8K photoreal, 180 degree shutter, pore-level skin, true fabric/hair physics; no 3D render, no plastic skin. NO IP / NO BRANDS / NO LOGO
影片模型 Kling 3.0（含 3.0 Turbo）｜微距一鏡＋聲音（OPTICS、FIRST FRAME 座標）
官方音訊段範例：FIRST FRAME 用百分比座標定位（錶盤佔畫面寬 80%，x 55%、y 50%），OPTICS 寫出起止距離（25cm → 3cm），AUDIO 只要近收音滴答聲、無音樂。
官方範例提示（英文原文）
SCENE CONTEXT A single continuous extreme macro push-in on a scratched vintage military field watch held in a man's fingers, from a full view of the dial into an extreme close-up of the aged hands at its center. FIRST FRAME AND BLOCKING Frame one shows the full watch filling 80% of frame width, dial at x 55%, y 50%, tilted slightly toward camera against a near-black void with faint warm falloff. Weathered fingers grip the case at frame left and lower right, out of focus. Crown points screen-right. Time reads 7:37, second hand near 6. Watch: unbranded 1970s military field watch. Matte black dial speckled with dust and micro-scratches under a domed crystal. Cream aged-lume Arabic numerals 1-12, railroad minute track, small broad-arrow mark below center. Cathedral hour hand with two rounded lume lobes, pencil minute hand, thin steel second hand. Brushed steel case, worn chipped bezel, knurled crown. FORMAT Single continuous take, no cuts, no speed ramps. OPTICS Macro probe-lens character, 18 degree field of view narrowing into true macro. Camera starts 25 cm from the dial, ends 3 cm from the hands. Razor-thin focus tracks forward, keeping the hands sharp while numerals and case melt into warm bokeh; the final frame resolves lume grain and dust fibers. CAMERA Slow hypnotic dolly-in along the lens axis toward dial center, constant velocity, no drift, no reframing; precision-slider smoothness, not handheld. The hands' crossing point settles just right of center, flanked by the 8 numeral and the broad arrow. ACTION TIMING 0.0-2.0s: full dial in frame, fingers gripping; second hand ticks; push-in begins. 2.0-4.5s: fingers exit frame edges as magnification grows; numerals slide past; focus tracks the hands. 4.5-6.0s: extreme macro on the layered hands over the black dial; dust and scratches fully resolved; second hand keeps moving to the last frame. PHYSICS Second hand advances with real mechanical cadence. Faint micro-tremor from the holding hand early on. Dust motes drift through the light near the crystal. No morphing of numerals or hands. LIGHTING Warm tungsten-amber key from upper left, raking low so every scratch and embossed numeral casts a tiny shadow. Deep low-key: background crushed near-black, highlights only on case rim, hand edges and lume. Golden patina palette, no flat front light. AUDIO Soft close-mic ticking, faint room tone. No music. LOCKS No logos or readable text besides numerals and minute track. Photoreal macro, fine film grain.
影片模型 Kling 3.0（含 3.0 Turbo）｜CLI：首幀＋pro 模式
官方 README 範例（譯：黎明時鏡頭緩慢推入森林空地）。
官方範例提示（英文原文）
higgsfield generate create kling3_0 \ --prompt "slow camera push through a forest clearing at dawn" \ --start-image ./first.png \ --duration 5 --mode pro --sound off \ --wait
影片模型 Kling Motion Control 3.0（及 2.6）｜文字只改環境，動作交給參考影片
官方 2.6 指南示範（譯：一隻柯基跑進來，繞著女孩的腳…）——不受限於參考影片的背景，可用文字改變環境、加入元素。
官方範例提示（英文原文）
A corgi runs in, circling around a girl's feet
影片模型 Google Veo 3.1（含 Lite、Veo 3）｜Multi Reference Mode：一句提示 → 最多 4 個連接場景
官方：寫一個簡單提示，Veo 3.1 Standard 自動生成最多四個互相連接、轉場流暢的場景。適合要「一組」連貫鏡頭而不想逐鏡寫的情況。
影片模型 Gemini Omni Flash（含 1.1）｜結構化基本提示（標註參考輸入）
官方範例（譯：@character_photo 站在黃昏雨濕的市集攤檔前，直視鏡頭，緩慢推軌，攤檔的暖鎢絲燈，淺景深）。主體＋動作 → 場景 → 鏡頭 → 光線。
官方範例提示（英文原文）
@character_photo standing at a rain-slicked market stall at dusk, looking directly at camera, slow dolly in, warm tungsten light from the stall, shallow depth of field.
影片模型 Gemini Omni Flash（含 1.1）｜品牌約束：把限制直接寫明
官方建議的品牌廣告約束寫法（原文引號內句子）。
官方範例提示（英文原文）
The label color stays consistent across all cuts. The product shape does not distort.
影片模型 Gemini Omni Flash（含 1.1）｜真人＋3D CGI 混合
官方範例：手按鍵盤，藍色毛茸卡通角色從螢幕跳出。精確描述位置、光線、鏡頭與動作，模型一次過產出。
官方範例提示（英文原文）
A cinematic mix of live-action and 3D CGI. A hand presses a key on a laptop, and a small, blue, furry cartoon character wearing pink glasses and a Hawaiian shirt jumps out of the screen. The character multiplies rapidly, flooding the desk and then a busy city intersection. A yellow taxi hits one of the small creatures, causing it to magically grow into a giant, Kaiju-sized monster. The giant blue furry creature stomps through the city streets, pressing its huge hands against a glass office building while people run away. A young Asian man in a purple hoodie stands on the street, pulls out a glowing pink and green box, and opens it. A magical pink light beams out, sucking the giant monster and all the small creatures into the box like a vacuum. Cinematic lighting, dynamic camera angles, blockbuster movie VFX.
影片模型 Gemini Omni Flash（含 1.1）｜「保留一切，只換角色」編輯
官方範例：先逐項列出要保留的（靜態中景、白色辦公桌、黑色旋轉電話…），再寫「Replace the character entirely with a new one」並詳述新角色。先鎖定、後替換。
官方範例提示（英文原文）
Preserve everything about the scene exactly as it is: the static medium shot of a character sitting behind a white office desk, waist-up framing, in front of a large floor-to-ceiling window overlooking a dense sunny city skyline. Preserve the desk setup untouched, the black rotary telephone, the folded newspaper with a yellow pencil resting on it, the small black leather notebook, and the desk pad. Preserve the exact timing and actions: the character holds the phone receiver to their ear from the start, hangs it up at the same moment, then turns toward the window, exhales, and leans back with arms settling in a relaxed, satisfied posture. Preserve the camera, which stays completely static throughout, the bright natural daylight pouring in from the window, and the soft shadows across the office. Preserve the audio exactly the same spoken line, the same phone-hang-up sound, and the same exhale. Replace the character entirely with a new one: swap the previous figure for a young man with dark curly hair and a chiseled jawline, wearing an oversized tan/beige retro 1980s business suit over a light blue button-up shirt, finished with a bold leopard-print silk tie. Give him the same confident, corporate energy, focused and slightly tense while on the call, then relieved and smirking as he leans back. Every other element of the shot, motion, timing, lighting, and sound must remain identical to the original.
影片模型 Gemini Omni Flash（含 1.1）｜VFX 加元素：SOURCE LOCK（V2V）
官方 VFX 範例：開頭聲明「This is an ADD-ELEMENT VFX pass on video_1, NOT a new generation. Keep 1:1 every…」——明確告訴模型這是加元素，不是重新生成。
官方範例提示（英文原文）
=== BIRD LEAVES THE SCREEN — V2V on video_1 (~10s, 16:9, 30fps) === SOURCE LOCK: This is an ADD-ELEMENT VFX pass on video_1, NOT a new generation. Keep 1:1 everything already in the plate—the black monitor, the desk setup, the hand and its motion/timing, and the exact CAMERA path (the initial hold, the pan to the right, and the slow sweep across the office space) and the EDIT. Do NOT re-frame, re-time, re-angle, re-cut, or change the hand's positioning. Only ADD the live bird and animate it. SUBJECT: One real Common Kingfisher—bright blue plumage on the back, vibrant orange-rufous underparts, long black bill, short red legs. It must 100% match the appearance and proportions of the kingfisher shown on the monitor screen. THE MECHANIC (mapped onto the plate's own beats): 0–1.5s: The bird stays perched on the branch inside the monitor image (screen image unchanged), then subtly comes alive—a quick head twitch, blink, and slight breathing movement. 1.5–2.5s: As the hand waits, the bird leans forward and flutters out of the display plane, crossing from the flat screen into real 3D space in front of the monitor with rapid, sharp wingbeats. Behind it, the image on the monitor screen seamlessly shows the same branch but empty. 2.5–3.5s: The bird alights on the open palm—real weight settles, feet grip the hand, wings fold tight, it glances around. 3.5–5.5s: As the camera begins to pan to the right, the bird stays perched on the hand, tracking with the hand's motion through the frame. 5.5–7s: As the hand remains extended and the camera continues scanning the office, the bird crouches and launches with a sharp, fast downstroke, taking off into the open office space. 7–10s: The camera completes its pan across the office desks, partitions, and curtains. The bird is seen flying dynamically through this background 3D space, darting between the workstations before exiting the frame or fading into the distance near the background curtains. LIGHT-MATCH / INTEGRATION (top priority): The real bird is lit by the room—the bright overhead LED panel light and ambient warm office lights. It must reflect the same exposure, color grade, and grain as the plate, not the outdoor lighting from the original screen image. Precise soft contact shadows of the bird and its feet must cast onto the palm. Feathers should catch the cool overhead glare as it moves. ANTI-SLOP: Real feather texture with high detail; lively eyes with a sharp catchlight; convincing flight physics with proper weight, acceleration, and air resistance. No robotic/CGI look, no floating, and no morphing shapes or extra limbs.FORBIDDEN: Changing the original camera pan, altering the hand's position, modifying the office background, or changing the timing of the camera movement.Diegetic SFX only: Rapid, high-pitched wingbeats, a sharp kingfisher whistle chirp, and the ambient office hum from the plate. No music. No on-screen text.
影片模型 Cinema Studio（4.0 / 3.5 / 3.0 及早期版本）｜4.0 Era 選擇：同一場景設定為 2010s
官方範例：同一角色、同一編舞、同一被拒絕橋段，只改年代。注意 BEAT 分段、CHARACTERS 要求配對參考圖並換上 2010s 服裝、AUDIO 要求「original, fictional … (not any existing song)」避免版權。
官方範例提示（英文原文）
15-second multi-shot cinematic scene in the style of a 2010 dance movie. Shot on a digital cinema camera: crisp clean image, deep blacks, glossy teal-and-orange color grade, smooth steadicam movement, 120fps slow-motion accents, rhythmic cuts landing on the beat, streaky lens flares from lasers. Location: a 2010s nightclub, pulsing LED video walls, sweeping laser beams, strobes, thick haze from a fog machine, an illuminated bar glowing in the background, a DJ on a raised booth behind CDJs and a laptop. CHARACTERS (match the faces and hairstyles of the attached reference images exactly, but dress them in 2010s club fashion; keep them consistent across all shots): GOLD GIRL: young woman, voluminous brunette blowout, large gold hoop earrings; wearing a gold sequin bodycon mini dress and black strappy heels. PINK GIRL: young woman, short brunette bob with bangs; wearing a hot-pink bandage mini dress, a chunky silver statement necklace and metallic heels. THE GUY: young man with a large afro, thin mustache and soul patch; wearing a slim-fit light-blue blazer over a deep V-neck white tee, dark skinny jeans, crisp white sneakers and dark wayfarer-style sunglasses indoors. A packed crowd of 2010s-dressed extras surrounds the floor, hands in the air, a few filming on early smartphones and compact digital cameras, all bouncing to the beat. BEAT 1, THE CLUB (0-3s): Steadicam glides through the hazy crowd, laser beams cutting overhead, LED wall pulsing; the crowd parts to reveal GOLD GIRL and PINK GIRL dancing side by side in perfect sync at the center of the floor, owning it. Extras cheer and mirror their moves. BEAT 2, THE ENTRANCE (3-6s): Cut to THE GUY sliding through the crowd; he lowers his sunglasses down his nose, a brief 120fps slow-motion hero moment with a laser flare across the lens, then a speed-ramp back to real time as he struts toward GOLD GIRL doing an exaggerated shuffle step. BEAT 3, FIRST ATTEMPT (6-9s): Quick cuts on the beat. He orbits GOLD GIRL, mirroring her moves, leaning in with a smug grin and offering his hand mid-move. Snap to her deadpan face: she raises a flat palm in a "stop right there" gesture, spins away without missing a step. Crowd goes "oooh". BEAT 4, SECOND ATTEMPT (9-12s): Undeterred, he shuffle-glides across the floor to PINK GIRL, drops a flashy pose and wiggles his eyebrows over his sunglasses. She looks him up and down, flips her hair dismissively and struts off to join GOLD GIRL. A DJ air horn blares from the booth. BEAT 5, DOUBLE REJECTION (12-15s): Wide shot: the girls fist-bump and keep dancing back to back, untouchable. THE GUY shrugs an exaggerated "well, I tried" shrug straight to camera as confetti cannons pop in slow motion around him, glittering in the lasers. The crowd bursts out laughing and keeps jumping; camera cranes up over the flashing LED wall. AUDIO: An original, fictional electro-house club banger (not any existing song), four-on-the-floor kick at 128 BPM, side-chained pumping synths, rising build into a big drop at the entrance of THE GUY, chopped wordless vocal hooks, glossy 2010 radio-dance mix. Crowd cheers and claps on the beat. Beat 3: crowd gasps "oooh!". Beat 4: a loud DJ air horn as PINK GIRL rejects him. Beat 5: the drop hits full force under crowd laughter and cheers as the confetti flies. No spoken dialogue.
影片模型 Cinema Studio（4.0 / 3.5 / 3.0 及早期版本）｜4.0 Era 選擇：同一場景設定為 1980s
與上例對照：只改年代描述（35mm 菲林顆粒、光暈、片門晃動、早期 80 年代院線拷貝質感）及服裝，其餘結構相同。
官方範例提示（英文原文）
15-second multi-shot cinematic scene, 1980s dance-movie style. Shot on 35mm film: heavy visible film grain, halation blooming around practical lights, subtle gate weave, warm faded highlights, slight chromatic softness, the texture of an early-80s theatrical print. Location: a retro discotheque with a glowing red, white and blue checkerboard illuminated dance floor, chrome railings, red tufted leather booths, mirror balls and tinsel garlands overhead, starburst lens flares from the rig lights. CHARACTERS (match the attached reference images exactly, keep faces, hair and outfits consistent across all shots): GOLD GIRL: young woman, voluminous brunette 70s blowout, gold sequin sleeveless top over a mustard long-sleeve, orange high-waisted flared pants, gold platform shoes, large gold hoop earrings. PINK GIRL: young woman, short brunette bob with bangs, hot-pink satin top trimmed with pink marabou feathers, silver sequin mini skirt, long pink gloves with feather cuffs, leopard-print boots. THE GUY: young man with a large afro, thin mustache and soul patch, amber-tinted aviator glasses, light-blue suit jacket over a cream wide-collar shirt, black flared trousers, gold hoop earring. A background crowd of era-dressed extras rings the dance floor behind the chrome railings, watching, clapping on the beat and mirroring the moves. BEAT 1, THE FLOOR (0-3s): Crane shot sweeps down from the mirror balls toward the glowing floor, then a fast crash zoom onto GOLD GIRL and PINK GIRL dancing side by side in perfect sync at the center, hips and shoulders hitting every beat. The crowd behind the rails claps along. BEAT 2, THE ENTRANCE (3-6s): Whip pan to THE GUY sliding into frame on his knees across the lit tiles, popping up into a confident funky strut. Low-angle dolly follows him as he points finger-guns at the girls. Extras hoot and cheer. BEAT 3, FIRST ATTEMPT (6-9s): Quick cuts on the beat. He spins around GOLD GIRL, mirroring her moves, leaning in with a smug grin and offering his hand mid-move. Snap zoom to her face: she raises an eyebrow, spins out of his reach and turns her back on him without missing a step. Crowd goes "oooh". BEAT 4, SECOND ATTEMPT (9-12s): Undeterred, he moonwalk-glides over to PINK GIRL, drops a flashy pose, lowers his aviators and winks. Snap zoom to her deadpan face; she flips her feather-trimmed glove in a dismissive "talk to the hand" gesture and dances away to join GOLD GIRL. BEAT 5, DOUBLE REJECTION (12-15s): Wide shot: both girls stand side by side, arms crossed, shaking their heads in comic unison while still bouncing to the beat. THE GUY shrugs an exaggerated "well, I tried" shrug to the camera. The surrounding crowd bursts out laughing and keeps dancing. Freeze-frame on his sheepish grin, film grain heavy, image flickers like a projected print, classic 80s comedy ending. AUDIO: An original, fictional upbeat funk-disco instrumental (not any existing song), punchy slap bass, four-on-the-floor drums, bright brass stabs, wah-wah guitar, around 118 BPM, mixed like a vintage vinyl record with warm tape saturation. Crowd claps land on the beat. Beat 3: crowd gasps "oooh!". Beat 4: a short vinyl record-scratch as PINK GIRL rejects him. Beat 5: warm crowd laughter and cheers; the music hits a triumphant brass sting on the freeze-frame. No spoken dialogue. Step 6. Use Color Grading as a final pass. After generation, Color Grading gives you post-production control without leaving the platform. Adjust temperature, contrast, saturation, sharpness, film grain, highlights, and exposure to finish the clip.
影片模型 Cinema Studio（4.0 / 3.5 / 3.0 及早期版本）｜4.0 Emotion Wheel：在提示中標記角色情緒
官方語法：`@character_name` 後接情緒（Joy、Anger、Fear…），模型據此生成表情與肢體語言。
模板
@character_name Joy
影片模型 Cinema Studio（4.0 / 3.5 / 3.0 及早期版本）｜3.5 設定組合：戰爭場面（只靠面板）
官方三個實戰設定之一（15 秒海上登陸）。Bleach Bypass＝高反差、銀灰去飽和；Raw Chaos＝貼近爆炸的劇烈晃動；Practicals＝場內光源。
影片模型 Cinema Studio（4.0 / 3.5 / 3.0 及早期版本）｜3.5 設定組合：拉力賽追逐
官方設定：Cold Steel 色板＋Anamorphic＋14mm 超廣角，營造長片級追逐。
影片模型 Cinema Studio（4.0 / 3.5 / 3.0 及早期版本）｜3.5 設定組合：黑暗奇幻戰役
官方設定：Epic＋Teal & Orange Epic＋Epic Scale 鏡頭＋Contre-jour 逆光，令奇幻戰役讀起來像真人實拍。
影片模型 Cinema Studio（4.0 / 3.5 / 3.0 及早期版本）｜3.0：留空提示，只選角色與地點
官方 3.0 教學：角色（Cast）與地點設定好後，可以完全不寫提示。
官方範例提示（英文原文）
Leave the prompt empty. Select your character and your location. Cinema Studio handles the rest.
影片模型 Cinema Studio（4.0 / 3.5 / 3.0 及早期版本）｜3.0：指明每張圖的角色（Image 1 = 角色、Image 2 = 地點）＋多鏡打鬥
官方範例：首句分配參考圖用途，之後 Multi-shot editing 逐鏡描述酒店走廊打鬥。
官方範例提示（英文原文）
Use Image 1 as a reference for the character's appearance, and Image 2 as reference for the location. Multi-shot editing, hotel corridor fight sequence. Shot 1: Telephoto lens. At the far end of a long, narrow luxury hotel corridor stands a young man in a dark tailored suit, white dress shirt, and a slightly loosened tie. The corridor features a deep red carpet, beige walls with warm yellow wall sconces, and evenly spaced room doors on both sides. Blocking his path are five or six men in black suits. They unbutton their jackets and assume fighting stances. Shot 2: Handheld tracking shot. The protagonist strides quickly toward the first opponent. In the tight space, he slams an elbow into him, smashing him against the wall. A framed picture rattles loose and shatters on the floor. Shot 3: Low-angle side shot. A second attacker throws a punch from the side. The protagonist grabs the arm mid-swing, uses the corridor wall for leverage, and flips him hard onto the carpet. Shot 4: Fast frontal close-up. The protagonist slips a punch and counters with a sharp palm strike to the jaw. The attacker crashes backward through a hotel room door, falling inside. Shot 5: Handheld, shaky tracking shot mid-hallway. Two attackers close in simultaneously. In the extremely confined space, the protagonist blocks left and right, drives his knee into one man's abdomen, then elbows the other in the temple. Both collapse in succession. A wall sconce shatters during the scuffle, darkening part of the corridor. Shot 6: Dynamic tracking shot. The final attacker, the largest of them all, stands guard near the door at the end of the hall. The protagonist accelerates into a sprint. In the narrow space, he sidesteps a heavy punch and spins into a rotating elbow strike, knocking the man down hard onto the carpet. Throughout the fight, the suit remains largely intact, the tie slightly crooked, the shirt subtly loosened, a faint sheen of sweat on his forehead, but he remains composed and sharp. Shot 7: Static camera. Medium shot widening to full. At the end of the corridor, the protagonist calmly walks past the fallen men. He adjusts his tie with one hand, his steps steady and controlled. He stops before the final door at the far end, back to camera. His broad-shouldered silhouette is outlined by the warm light of the wall lamp above the door. He slightly turns his head, takes a slow breath, and raises his right hand toward the door handle. Freeze frame the moment before his hand touches the handle.
影片模型 Cinema Studio（4.0 / 3.5 / 3.0 及早期版本）｜3.0：延續 @video1（無需新關鍵幀）
官方範例：以「continuing @video1」接續上一段，逐鏡標明景別與內外景。
官方範例提示（英文原文）
Multi-shot cinematic scene, continuing @video1 Shot 1 (Wide, interior): Luxury high-rise hotel suite. Floor-to-ceiling windows reveal a city skyline under an orange-gold dusk sky. Interior dim, warm tungsten lighting. White bed and bedside lamp visible. Soft foreground bokeh. Light rain marks on the glass. Camera faces the door. The door opens. A young man in a slightly disheveled suit steps in. He walks toward camera; the door closes behind him. Near the windows stands a woman holding a glass of red wine, silhouetted by the sunset. He passes camera, heading toward her, hurriedly adjusting his jacket, straightening his tie, smoothing his shirt. Shot 2 (Medium tracking): He discreetly breathes into his hand, smells it, frowns. Nervous. He composes himself and continues forward. Man: "Sorry, I got held up a little. I should've been here sooner." Shot 3 (Over-the-shoulder from behind the woman): She slowly turns, gently swirling her wine, studying him with a faint smile. Woman: "Honey, where have you been?" Shot 4 (Two-shot, medium close-up): He steps close and wraps his arms around her waist from behind, forcing tenderness. Man (evasive): "I... was jogging." Shot 5 (Close-up on woman): She glances down at his shirt, raises an eyebrow. Woman: "That's weird... your shirt is completely dry." A brief pause.
影片模型 Cinema Studio（4.0 / 3.5 / 3.0 及早期版本）｜3.0：純文字生影（大廈爆破）
官方純文字範例：事件＋物理過程（逐層向內塌陷、玻璃與混凝土塵）。
官方範例提示（英文原文）
A demolition crew detonates a skyscraper in a dense city at dawn. The building folds inward floor by floor in a cascade of glass and concrete dust, debris clouds billow outward in slow motion, pigeons scatter across the orange sky. Shot from street level on a handheld camera behind a safety barrier, crowd reactions visible in foreground. Realistic shockwave dust rolling toward camera.
影片模型 Cinema Studio（4.0 / 3.5 / 3.0 及早期版本）｜3.0：換掉現有影片的地點
官方範例：極短指令——在 @video 中把地點換成 @image_1，加類型與動作。
官方範例提示（英文原文）
In @video change location to @image_1. Horror film, man running from something scary.
影片模型 Cinema Studio（4.0 / 3.5 / 3.0 及早期版本）｜3.0：產品廣告（時間碼分段）
官方薄餅廣告範例：用「0–4 seconds:」等時間碼分段敘事。
官方範例提示（英文原文）
0–4 seconds: The guy from @image_1 shuffles sleepily into the kitchen in pajamas, hair messy, eyes half-closed. He pulls the fridge open — completely empty. He stares blankly, sighs. He pulls out his phone. Cut to close-up of the phone screen — a food delivery app is open, a large bright button reads "Order Food". His thumb hovers, then taps it decisively. Camera stays tight on the screen, button animates as pressed. No scrolling, no other interactions. 4–9 seconds: Cut to exterior — bright daylight, modern apartment building. A delivery drone flies into frame from above carrying a pizza box. It descends in a smooth cinematic arc toward the balcony, gently places the pizza box on the railing, and lifts away. Wide-angle, slightly heroic drone shot, clean blue sky background. 9–15 seconds: Packshot. The guy is positioned on the left side of the frame, holding an open pizza box, taking a satisfied bite, eyes closed with joy. The right side of the frame is intentionally left clean and empty, neutral background, no objects, no clutter. Warm soft lighting. Static composition, camera does not move. Close-up on face left, right third of frame completely open for logo placement.
影片模型 Cinema Studio（4.0 / 3.5 / 3.0 及早期版本）｜3.0：Logo 動畫（Liquid Glass）
官方動態設計範例：指定整體美學（Liquid Glass）再列出詳細要求。
官方範例提示（英文原文）
Create an elegant dynamic logo reveal animation for the logo in Image 1. The overall visual style should follow a Liquid Glass aesthetic. Detailed requirements: The logo material should present a highly transparent liquid glass texture, featuring realistic refraction, transmission, and reflection effects. Inside the glass, flowing liquid light and subtle air bubbles move organically. The edges display characteristic glass dispersion, splitting light into a subtle rainbow spectrum. The logo should emerge from nothing through a transformation process where liquid glass material floats and flows in midair, then gradually solidifies into the final logo shape. Surface tension effects must be visibly realistic during the formation process. The overall animation rhythm should be extremely dynamic and powerful. The camera should switch angles multiple times throughout the sequence — from extreme close-up detail shots to full wide shots, from low-angle to high-angle views, including 360-degree orbital rotations around the logo. The camera movement enhances spatial depth and dimensionality, showcasing how refraction and internal light flow change from different perspectives. Transitions between shots should be smooth and fluid. Light passing through the glass logo should create caustic effects and colorful projections. The pacing should feel impactful and energetic, supported by fast, precise editing. The background should remain clean and dark to emphasize the glass material. At the end of the animation, the logo stabilizes in the center of the frame, with the glass surface maintaining a subtle, breathing-like liquid motion loop.
影片模型 Cinema Studio（4.0 / 3.5 / 3.0 及早期版本）｜3.0：2D 手繪動畫動作場面
官方範例：第一句定畫風與製作水準，再寫超現實城市變形動作。
官方範例提示（英文原文）
2D hand-drawn anime style with top-tier cinematic production quality. A surreal urban transformation action scene set in a modern city at night, with neon lights reflecting on rain-soaked streets. Suddenly, the entire city begins to slowly move and reorganize under the control of someone's psychic power. Giant skyscrapers rotate and tilt like building blocks. Roads rise from horizontal to vertical, transforming into massive walls, while former walls flip over to become new floors. The direction of gravity constantly shifts. Two warriors continue an intense, relentless battle within this continuously transforming geometric space. One hangs completely upside down in midair, launching a fierce attack toward the opponent below. The other uses a moving building façade as a springboard to propel upward in counterattack. Both fighters instantly perceive and exploit the evolving spatial structure around them. An impossible-architecture aesthetic runs throughout the scene: staircases lead into the void, corridors bend into loops, and building cross-sections simultaneously display contradictory perspectives from multiple angles. The camera rotates and rolls at extreme speed, following the two warriors through the chaotic environment, rapidly switching from overhead shots to low-angle views to side perspectives. Fragments of buildings float in the air, slowly rotating like a frozen asteroid belt. Shattered neon tubes scatter pink and blue glowing particles drifting throughout the space. Glass curtain walls explode, with shards floating in zero gravity. The warriors' clothing and hair are pulled simultaneously by gravitational forces from multiple directions. The space feels extremely chaotic yet internally coherent, creating a psychedelic and awe-inspiring visual impact. Cel-shaded coloring style, hand-drawn line texture, strong depth-of-field effects, and cinematic widescreen composition.
影片模型 Marketing Studio（影片）｜App 廣告：完整用戶旅程（UI 圖靜態顯示在手機上）
官方 CUE 系列：每張參考圖都說明「是甚麼」及「何時靜態顯示在手機螢幕」。目標人物：回家很累、不知煮甚麼的人——問題、掃描、食譜、烹調、成果。
官方範例提示（英文原文）
@image_1 is the CUE app scan screen — display it statically on the phone screen when scanning ingredients. @image_2 is the CUE app result screen — display it statically on the phone screen after scanning. Do not animate any app interface. A woman stands in a bright modern kitchen facing the counter. Spread across the counter in front of her: eggs, tomatoes, spinach, cheese, onion — random, loosely arranged. She stares at them. Arms loosely crossed. Thinking. She says out loud in English: "What should I cook?" A beat of silence. Then her face shifts — eyes widen slightly, one finger raises straight up in the air. The idea has arrived. She grabs her smartphone off the counter. She holds the phone out and aims it directly at the ingredients laid out on the counter in front of her — phone screen shows @image_1 static CUE scan UI, a scanning frame locking onto the ingredients below. A brief pause — then @image_2 static CUE result screen appears on the phone. A recipe. She reads it, nods slowly, one corner of her mouth pulls into a small smile. She sets the phone down. She turns to the counter with new purpose. Fast cuts: hands sorting ingredients into clean groups — eggs to one side, tomatoes grouped, spinach stacked, cheese pulled out. Deliberate, efficient, energized. The counter goes from random to organized in seconds. Cooking begins — knife through tomatoes, eggs cracking into a pan, spinach wilting in butter, cheese grated over the top. Steam rises. Sizzle fills the frame. She moves with calm confidence, referencing the phone screen once mid-cook with a quick glance. Final shot: She picks up the plate with both hands, turns to face the camera directly. Quiet satisfaction on her face. She looks at the dish, then back at the lens. Says "Use CUE" Camera: mix of medium shots, close-ups of hands and ingredients, over-the-shoulder during phone scan, fast cuts during prep, slow final hold on her face and plate. Lighting: bright soft natural kitchen light, warm tones, clean shadows, golden steam on final plate shot. Sound design: ambient kitchen — ingredients placed on counter, phone tap, shutter click, soft notification chime, fast knife cuts, pan sizzle, butter pop, steam.
影片模型 Marketing Studio（影片）｜第二個客群：健身用戶（同產品、不同鈎子）
官方範例：服裝強調「completely blank fabric, no logos, no brand marks」避免商標。
官方範例提示（英文原文）
A fit muscular young man in @image_1 in a plain black gym stringer, completely blank fabric, no logos, no brand marks, no prints, no text, no graphics on clothing whatsoever. Sweaty after workout, opens a messy fridge packed with random scattered ingredients — chicken, eggs, random vegetables, sauce bottles, everything chaotic and unorganized. He looks confused, scratches his head. He pulls out the ingredients one by one — chicken breast, eggs, bell pepper, avocado — and places them on the kitchen counter in a messy pile. He pulls out his phone, holds it above the ingredients and takes a photo — camera shows the phone screen capturing the food from above, the app interface appears on screen with the scanned ingredients @image_2 — app screen reference, static, no tapping). The phone screen fills the frame showing a complete recipe with macros generated automatically by the app. He nods confidently. Fast-paced cooking montage: overhead knife chopping with precision, oil hitting a hot pan in slow motion, chicken sizzling perfectly, vegetables tossed in the air and caught in the pan. Final shot: he sits at the counter with a perfect plated high-protein meal, takes a bite, gives a chef's kiss to camera, phone with the app propped up next to the plate @image_2 — app screen visible in background). A bold white caption text overlay appears centered in the lower third of the frame reading "POV: You're a gym bro who doesn't know how to cook..." in a clean sans-serif font, white text with subtle drop shadow against the video.
影片模型 Marketing Studio（影片）｜更電影感的 UGC：連續 360° 環繞
官方範例：同一產品故事，以連續慢速 360 度環繞鏡頭把整個烹調壓縮成優雅生活廣告。
官方範例提示（英文原文）
@image_1 is the phone camera UI — display it statically on the phone screen when the character photographs the ingredients. @image_2 is the CUE app screen — display it statically on the phone screen after uploading the photo. Do not animate any app interface. A man stands in a bright, modern kitchen with white countertops, warm natural light, and clean minimalist decor. 0–3s: Slow 360-degree orbital camera begins moving around the kitchen — smooth, cinematic, continuous rotation. The man opens the refrigerator door. He reaches in and pulls out ingredients one by one: tomatoes, eggs, a block of cheese, fresh greens. 3–6s: Orbital camera continues its slow rotation around him. He places each ingredient neatly onto the counter, arranging them with calm intention. Ingredients clearly visible, neatly spaced on white countertop. He steps back, looks at the spread. 6–8s: He picks up his smartphone. Over-the-shoulder angle — he points the phone DIRECTLY DOWN at the ingredients on the counter. Phone screen shows @image_1 — static camera UI. He taps. Shutter click sound. Then phone screen switches to @image_2 — static CUE app screen. He glances at it, nods once. 8–12s: Timelapse sequence — the orbital camera keeps its slow 360 rotation while cooking happens in accelerated time: knife slicing tomatoes, eggs cracking into a hot pan, cheese grated, greens tossed in. Steam rising. Pan sizzling. Hands moving efficiently. The kitchen fills with warmth and motion. The orbital shot compresses the full cook into a few elegant seconds. 12–14s: Timelapse ends. Camera slows to real time. He plates the dish — one clean, confident motion onto a white plate. 14–15s: He turns directly toward camera, holds the beautifully plated dish up with both hands, smiles wide, and says out loud: "Use CUE." Camera: continuous slow 360-degree orbital rotation throughout — smooth, steady, cinematic dolly movement around the subject. Never stops moving until the final direct-to-camera hold. Lighting: bright soft natural daylight, warm kitchen tones, clean shadows, golden highlights on food. Sound design: original groovy instrumental hip-hop beat — upbeat, warm, rhythmic. Sound effects: fridge door open, ingredients placed on counter, phone shutter click, soft chime, knife chopping, pan sizzle, timelapse whoosh transition. Spoken words at end: "Use CUE." No voiceover, no background song from existing artists. IMPORTANT: CUE is a digital mobile app only — on phone screen only, no physical product. App interfaces are displayed as static screens — no animation, no UI transitions. Phone always aimed at ingredients on counter, never at floor or empty space. Style: super realistic, cinematic, lifestyle commercial, warm tones, smooth orbital cinematography, 4K.
影片模型 Marketing Studio（影片）｜CGI 級產品揭示（9:16）
官方範例：手機衝入虛空、冰塊碎裂露出食材、食材變成成品菜式，全部扣回 CUE UI。
官方範例提示（英文原文）
Vertical 9:16 cinematic shot. A modern smartphone floats in pitch-black void with subtle dark green gradient, displaying the CUE app interface — dark UI with glowing mint-green accents, "Cook with what you have" headline visible, suggestion chip "eggs + tomato → omelette" highlighted. A finger taps the screen and camera crash-zooms forward THROUGH the phone display into deep space. A translucent ice cube emerges from the darkness, mint-green rim light pulsing along its edges — trapped inside: a fresh egg and a cherry tomato with green spinach leaves. Sudden hyper-speed shatter: the cube explodes outward in extreme slow motion, crystalline shards flying past the camera with mint-green light streaks and frozen mist. Mid-air transformation in slow motion: the egg cracks open and golden yolk pours, the tomato splits into perfect dices, butter cubes melt, fresh herbs scatter, all spiraling through dark void with motion-blur trails. Camera orbits as ingredients converge and snap into a perfect golden folded omelette on matte black ceramic plate, glossy and gently steaming, garnished with diced tomato and a basil leaf, soft mint-green underglow. Smooth pull-back reveals the smartphone again, now showing the finished omelette photo on screen with the glowing mint-green "Cook with what you have" headline. Cinematic premium aesthetic, deep blacks, hyperrealistic detail, fast-paced ad editing energy, multiple dynamic camera moves, 9:16 vertical.
影片模型 Marketing Studio（影片）｜16:9 生活廣告（角色貫穿全片）
官方範例：由冰箱到餐桌，溫暖、親切、電影感，適合 YouTube 投放。
官方範例提示（英文原文）
@image_1 — recipe detail screen reference: vertical dark-themed app showing the full Chicken Avocado Wrap page — dish photo at top, title, cook time 15 min, difficulty Easy, calories 480 kcal, ingredients list, step-by-step instructions. @image_2 — final dish reference: two halved chicken avocado wraps stacked on a dark ceramic plate, cross-section showing chicken, avocado, rice, spinach, tomato filling, lime wedges on the side. @image_3 - kitchen location reference: teal-green cabinets, large stainless steel fridge on the left, warm evening window light, dark stone countertops. She carries the ingredients from the fridge to the kitchen counter — container in one hand, avocado and spinach tucked against the other arm, walking calmly toward the counter. Easy, natural, no fuss. She sets everything down. The phone is propped vertically on a small stand showing @image_1 — ingredients list and numbered instructions visible on screen. She glances at the phone, then turns to the camera with that easy playful look and says: "The instructions are so detailed, even I can't mess this up." A beat — she gives the camera a slow smirk, reaches over, picks up a small piece of diced tomato from the cutting board and pops it in her mouth. Completely unbothered. Quick intimate cuts of the cooking process: tortilla laid flat, chicken strips placed, spinach and avocado layered in, tomato added, wrap rolled tightly, pressed down, sliced cleanly in half. The finished wrap on a dark ceramic plate exactly as in @image_2. She sets the plate down, looks at it, then back at the camera. Small satisfied pause. She picks up one half, holds it slightly toward the camera and says with a light grin: "Cue said fifteen minutes. It was fifteen minutes." She takes a bite. Handheld, warm kitchen light throughout, phone screen with @image_1 slightly visible in background during cooking. Cinematic TV commercial, 16:9, realistic. Sound: soft chopping and prep sounds during cooking, warm music holds through the cooking sequence and eases into a clean resolve on the final line and bite.
影片模型 MiniMax H3（Hailuo H3）｜超動感微距產品廣告（10 秒 5 鏡、速度斜坡）
官方範例：先定義「英雄產品」外觀，再聲明「PHOTOREALISM IS CRITICAL … not CGI」，再逐鏡寫時間碼（Shot 1 (0.0–1.2s)…）、Color grade、Audio（no melody）。
官方範例提示（英文原文）
The hero product: a real, physical high-top football boot "VYORA" — turquoise-to-deep-purple gradient upper with hot-pink lightning-crack graphics, dark purple laces, knitted sock collar fading from purple to bright red-pink, cyan brand emblem on the collar, white "VYORA" wordmark on the lateral side, translucent iridescent soleplate with hot-pink studs. Keep the boot shape, graphics, wordmark, and colors identical in every shot. PHOTOREALISM IS CRITICAL: this is real product photography footage, not CGI. The boot must look like a real manufactured shoe — visible knit fibers and loop texture on the collar, real stitching seams, micro-scratches and natural imperfections on the soleplate, fabric grain on the laces, matte and gloss zones exactly like real synthetic leather and TPU, physically accurate soft shadows and contact shadows, natural lens depth of field and subtle film grain. No plastic-toy look, no smooth 3D-render surfaces, no exaggerated glow on the shoe itself. Hyper-dynamic macro football boot commercial, 16:9, 10 seconds, aggressive rhythmic editing with hard cuts and constant speed-ramping — shots snap from ultra-slow-motion to sudden fast-forward bursts and back. Background: a bright, luminous cyan studio world with absolutely no black and no gray anywhere — a seamless vivid cyan-to-turquoise gradient sweep, evenly lit and hyper-saturated, with soft violet and hot-pink glow spots drifting across it in parallax and gentle light waves rippling through the cyan like light underwater; the glossy floor is bright cyan too, reflective like polished colored glass; the palette mirrors the boot itself — cyan, turquoise, violet, hot pink — fresh, electric, premium. Shot 1 (0.0–1.2s): extreme close-up of the knitted collar and cyan emblem; the camera whips in fast from off-frame and ramps into extreme slow motion, raking key light revealing every individual knit loop and fiber, soft violet-pink glow drifting in the bright cyan background bokeh. Shot 2 (1.2–2.4s): crash zoom straight into the iridescent soleplate held toward camera; speed-ramp — violently fast, then a near-freeze on the hot-pink studs, sharp specular reflections sliding across the glossy translucent plate exactly like real molded TPU, the plate picking up turquoise reflections from the set. Shot 3 (2.4–4.0s): low-angle FPV-style camera swing arcing under and around the boot standing on a low glossy cyan plinth — one continuous accelerating swoop from heel to toe that ramps down to slow motion exactly as the white "VYORA" wordmark crosses center frame, violet and pink rim light flaring along the silhouette, fine synthetic-leather texture and pink lightning cracks clearly visible, a real soft contact shadow anchoring the boot. Shot 4 (4.0–5.8s): the boot slams down onto the glossy bright-cyan floor in extreme slow motion — a burst of fine white mist erupts from under the studs on impact, real physics, weight and a slight settle-bounce; the camera does a tight accelerating 180° orbit around it as the mist hangs frozen, glowing turquoise in the light, then rushes away. Shot 5 (5.8–10.0s): single continuous closing shot — one single VYORA boot standing alone in perfect side profile on the glossy cyan mirror floor, toe pointing screen-left; the camera performs a fast pull-back that ramps down into a slow, smooth orbital drift and finally settles to a locked static hero framing with the boot centered; behind it the cyan-to-turquoise gradient glows brightly with slow waves of violet and hot-pink light sweeping across like silk, their reflections rippling over the mirror floor around the boot; thin wisps of white mist drift low, a soft overhead light pools gently on the boot; in the last second the camera is dead still, the mist settles, the background gives one gentle bright pink-violet wave. Exactly one boot in the frame — never two or three, no duplicates in reflections other than the natural floor mirror. Color grade: bright, airy, hyper-saturated cyan-turquoise base with violet and hot-pink accents, luminous shadows tinted cyan (no black, no gray anywhere in the frame), crisp whites on the wordmark, strong micro-contrast, premium athletic editorial finish with a photographic film-like texture. Audio: no melody — a deep pulsing drone underneath, punchy impact hits on every cut, time-stretch "vacuum" sound design on each speed-ramp, tactile knit and stud-click foley on macro shots, a big boom with mist hiss on the floor slam, airy whooshes on camera moves, then sudden dead silence on the final locked frame.
影片模型 MiniMax H3（Hailuo H3）｜指令式修正：只描述要改的
官方第 6 步：「Send an instruction describing only what should change.」錯一個細節不必從原提示重新生成。
影片模型 Minimax Hailuo 2.3｜短而具體的物理動詞（melt / ignite / morph）
官方範例之一（另有「Human transforming into werewolf under moonlight.」「Anime girl turning into a light spirit.」）。官方建議：保持提示短而物理化。
官方範例提示（英文原文）
Perfume bottle melting into gold.
影片模型 Wan 2.7 / 2.6（含 WAN Camera Control）｜鏡頭控制：把視覺與情緒意圖合成一句
官方香水廣告例（譯：晨光穿過玻璃，鏡頭緩慢環繞香水瓶，表面金色反光，背景淡淡薄霧）。光線＋鏡頭動作＋材質反應＋氛圍。
官方範例提示（英文原文）
soft morning light through glass, slow dolly around perfume bottle, golden reflections on surface, subtle haze in the background.
影片模型 Grok Imagine（影片）／Grok Video 1.5｜CLI：Grok Video 1.5 圖生影
官方 CLI 範例（譯：電影感手持鏡頭，霓虹雨夜街道）。
官方範例提示（英文原文）
higgsfield generate create grok_video_v15 --prompt "cinematic handheld shot, neon rainy street" --start-image ./image.png --duration 5 --resolution 720p --wait
影片模型 FLUX 3 Video｜一鏡到底＋SUBJECT LOCK／ENVIRONMENT LOCK
官方範例：小提琴手在白色圓形大廳。用「SUBJECT LOCK」（身體永不轉動、雙腳不離原位）、「ENVIRONMENT LOCK」（建築幾何固定）、「Color grade locked throughout」逐項鎖定，最後「Total: 7s / 1 shot / 16:9」。
官方範例提示（英文原文）
single continuous shot, one take no cuts, cinematic oner, cinematic lighting, photorealistic, 35mm film quality, professional color grading, sharp focus, high detail texture, film grain, depth of field mastery, smooth slow dolly A young woman plays violin inside a white circular rotunda — long wavy brown hair half-tied with a scarlet-red ribbon bow, its tails hanging down her back, a flowing soft pink dress with draped sleeves. She stands on a round white platform encircled by a ring of deep mirror-blue water, smooth white curved walls with arched niches around her, a vast circular opening overhead revealing vivid blue sky with soft white clouds. The violin is natural honey-brown wood. She plays with genuine musical focus and feeling throughout, eyes closed. SUBJECT LOCK: she remains facing the camera the entire shot — her body never turns, never rotates, never spins; feet planted in the same spot on the platform from first frame to last; only her bow arm, fingers, breathing and a gentle sway of her head move with the music; the dress hem and ribbon tails move only from her subtle motion, nothing else. ENVIRONMENT LOCK: the architecture is rigid and static — walls, arches, platform, water ring and oculus keep exactly the same geometry, position and proportions throughout the shot; no new rooms, no morphing walls, no shifting arches; the water stays calm with only faint ripples. Color grade locked throughout: clean white architecture with pale icy-blue shading, deep saturated blue only in the water ring, vivid blue sky with white clouds in the oculus, soft pink only on the dress, red only in the hair ribbon, warm natural wood only on the violin, soft natural skylight from above, no harsh highlights, no blown-out whites, restrained contrast, no color shift first frame to last. Single continuous shot 5s: Opening on her in medium shot — violin under chin, bow drawing across strings in natural playing motion, eyes closed. Camera performs ONE simple move only: a slow, straight dolly back along the ground, no pan, no orbit, no crane — the frame gradually widening from medium shot to medium-wide, revealing more of the platform, the water ring and the lower arches; the oculus edge just entering the top of frame by the end. The camera keeps her perfectly centered and frontal the whole way. She stays absorbed, eyes closed, playing continuously as the camera settles. Total: 7s / 1 shot / 16:9
影片模型 Virality Predictor（brain_activity）｜CLI：評估成品影片
官方 README 指令。不要把「分析這段影片」誤路由到生成模型。
模板
higgsfield generate create brain_activity --video ./ad.mp4 --wait
影片模型 影片工具與工作流（去背、Draw to Video、Reframe、Explainer）｜Draw to Video：草圖幀＋時間點＋一句指令
官方 README（譯：把外套變紅色）。--image 是 --sketch 的別名。
官方範例提示（英文原文）
higgsfield generate workflow draw_to_video \ --video ./source.mp4 \ --sketch ./frame.png \ --timestamp 3.2 \ --prompt "make the jacket red" \ --wait
影片模型 影片工具與工作流（去背、Draw to Video、Reframe、Explainer）｜解說影片 Video block 模板（統一 STYLE）
官方 Skill 模板：每段都貼上同一組 STYLE 描述＋同一張風格關鍵圖，AUDIO 只放環境聲／音樂（不要人聲），NEGATIVE 排除寫實、3D、對嘴、字幕。STYLE 描述一律以「non-photorealistic, illustrated, not a photo, no live-action, no realism」結尾。
官方範例提示（英文原文）
Block N
STYLE REFERENCE: Match the attached reference image EXACTLY. Replicate its look precisely: {STYLE tokens}. Every element below rendered in that identical style.
SCENE: {scene and one clear action matching Block N narration}.
MOTION: {camera move and animation behavior—slow push-in, drift, scale shock, hard contrast cut}.
AUDIO: {ambient SFX or music only—NO voice, dialogue, or narration}.
NEGATIVE: color drift, photorealism, 3D render, lip-sync, captions, on-screen text, logos, watermark{, plus style-specific bans}.
音訊模型 Seed Audio 1.0｜單一隔離音效
官方 Skill 範例（譯：短促、獨立的劍擊聲，乾聲錄音室錄製，無音樂）。「dry studio recording」指定空間、「no music」做隔離。
官方範例提示（英文原文）
higgsfield generate create seed_audio \ --prompt "short isolated sword impact, dry studio recording, no music" \ --wait \ --json
音訊模型 Seed Audio 1.0｜以參考音訊保持同一把聲、只改語氣
官方範例（譯：同一把聲，更平靜的語氣），配 --audio-references ./voice.wav。
官方範例提示（英文原文）
higgsfield generate create seed_audio \ --prompt "same voice, calmer delivery" \ --audio-references ./voice.wav \ --wait
音訊模型 Seed Audio 1.0｜預設／複製聲音說一句台詞
官方 CLI 範例。--voice_id 與 --voice_type 由 higgsfield voices list 取得（可為 preset 或複製聲音 element）。
官方範例提示（英文原文）
higgsfield generate create seed_audio --prompt "Welcome to the show!" --voice_type preset --voice_id <voice_id> --wait
音訊模型 Seed Audio 1.0｜環境聲
官方 Skill 範例（譯：電影感雨聲環境，遠處雷聲）。
官方範例提示（英文原文）
cinematic rain ambience with distant thunder
音訊模型 Seed Audio 1.0｜「先聲後畫」：音訊作為影片模型的參考輸入
官方反向流程：①先用 Seed Audio 1.0 生成對白／環境聲；②確認節奏與長度配合計劃的鏡頭；③把音訊作參考輸入接到 Cinema Studio、Seedance 2.0 或 Kling 3.0；④生成影片，動作有機會對上聲音；⑤若不同步，通常以同一音訊重生影片較快。
音訊模型 Eleven v3（ElevenLabs）｜行內情緒標籤（方括號）
Eleven v3 以「Inline emotion tags」控制語氣（官方音訊指南）。方括號情緒提示的實際寫法取自官方 LipSync 指南（「Add emotion cues in brackets where the tone should shift: [excited], [calm], [direct]」）；套用到 Eleven v3 的具體標籤清單未核實。
模板
[excited] <line> ... [calm] <line> ... [direct] <line>
官方範例提示（英文原文）
[excited], [calm], [direct]
音訊模型 Eleven v3（ElevenLabs）｜CLI：text2speech_v2 + elevenlabs
官方 CLI 範例。
官方範例提示（英文原文）
higgsfield generate create text2speech_v2 --prompt "Hello from Higgsfield" --variant elevenlabs --voice_type preset --voice_id <voice_id> --wait
音訊模型 Sonilo Music｜器樂循環（情緒＋樂器＋no vocals）
官方 Skill 範例（譯：器樂、舒適森林循環，溫暖馬林巴與柔和弦樂，無人聲），30 秒。
官方範例提示（英文原文）
higgsfield generate create sonilo_music \ --prompt "instrumental cozy forest loop, warm marimba and soft strings, no vocals" \ --duration 30 \ --wait \ --json
音訊模型 Sonilo Music｜短曲風描述
官方 CLI 範例（譯：電影感合成波音樂），12 秒。
官方範例提示（英文原文）
higgsfield generate create sonilo_music --prompt "cinematic synthwave track" --duration 12 --wait
音訊模型 Mirelo Text to Audio｜單一近距離音效
官方 Skill 範例（譯：一下沉重木門猛關，近距離收音，無環境聲），2 秒。
官方範例提示（英文原文）
higgsfield generate create mirelo_text_to_audio \ --prompt "single heavy wooden door slam, close microphone, no ambience" \ --duration 2 \ --wait \ --json
音訊模型 Inworld Text to Speech｜遊戲角色短台詞
官方範例（譯：閘門開了，快走！）。台詞要短，不阻礙遊戲；每個角色鎖定一把聲。
官方範例提示（英文原文）
higgsfield generate create inworld_text_to_speech \ --prompt "The gate is open. Move!" \ --voice "<exact voice value from model get>" \ --wait \ --json
音訊模型 聲音工具：Change Voice、Translate、Dubbing、LipSync｜LipSync：為「說」而寫的劇本＋方括號情緒提示
官方：正式書面語令表演機械；口語化語言提供更多表情線索。在語氣轉換處加方括號情緒提示。先生成有表現力的音訊；來源圖用帶輕微自然表情（不是證件照式中性臉）；用訓練過的身份層（如 Soul ID）保持多片一致。
官方範例提示（英文原文）
[excited], [calm], [direct]