APIMODELSAPIMODELS

Seedance 2.5 Action & Fight Prompts

Seedance 2.5 video prompts, each shown next to the clip it produced and credited to its author. Where the creator published their prompt we keep it verbatim; where they did not, we reverse-engineer one from the finished clip and label it as such. Hover to play, one click to copy.

← All 95 Seedance 2.5 prompts

Last updated: · Seedance 2.5 API — pricing & playground · API docs

TL;DR

Seedance 2.5 is ByteDance's video model, unveiled on 23 June 2026 at the Volcano Engine FORCE conference. On the shipped API it generates 4 to 30 seconds in a single pass — or duration -1 to let the model choose the length — and accepts up to 50 reference assets in one job: up to 30 images at 4K each, up to 10 video clips and up to 10 audio clips totalling 30 seconds each. It adds video editing and video extension as task types, and generates audio in the same latent space as the picture. Resolution is 480p or 720p; the native 4K at 10-bit shown at the FORCE launch is not exposed on the API. ByteDance also published an official prompting guide for it. This page collects Seedance 2.5 video prompts with the clip each one produced, free to copy. Since 8 August 2026 you can run Seedance 2.5 on apimodels.app as the model seedance-2.5, from $0.120/s; the creators in this library ran these prompts in Dreamina (Jimeng) and Higgsfield.

Source: ByteDance Seed (the model vendor’s official site)

Hover to preview · click for sound

Thirty seconds of hand-to-hand combat in a rain-flooded metro maintenance hall

生成一段完整30秒、16:9横屏、24fps、写实电影级质感的现代近身格斗视频。使用我上传的两组人物参考图:第一名人物固定为"主角",第二名人物固定为"敌人"。严格继承参考图中两人的面部、年龄、发型、体型、身高比例、服装、鞋子、配饰和整体气质。全程不得交换身份,不得变脸、改变服装颜色、改变体型或生成第三名参战者。 【核心风格】 参考《一个人的武林》所呈现的现代硬派武术电影气质:动作迅猛、贴身、凶狠,攻防转换极快,拳脚具有真实接触感和明确受力反馈,但不照搬电影中的人物、场景和具体镜头。以高质量ACT动作游戏的第三人称战斗镜头诠释两人对决,摄影机主要跟随主角,位于主角肩后、腰后或侧后方,使观众产生操控主角迎战强敌的沉浸感;关键格挡和重击时,可以短暂切入侧面中景或近景。 视频不要求真正一镜到底,可以在30秒内进行多次自然切镜,但整场战斗必须像一个连续长镜头般流畅。利用人物身体遮挡、立柱擦镜、快速摇镜、撞击震动和动作匹配完成隐形切镜。切镜后必须延续上一镜的动作惯性、人物位置、身体朝向、伤势和攻击方向,禁止瞬移、换位错误、人物突然恢复站姿或摄影机无理由越轴。 【场景】 深夜暴雨,一座已经停运的高架地铁检修站。站厅呈狭长矩形,地面是被雨水浸湿的深灰色防滑地砖,分布少量浅水洼,能够反射冷白顶灯和红色维修警示灯。左侧是一排金属检票闸机和关闭的卷帘门,右侧是半透明钢化玻璃护栏,护栏外能够看到高架轨道、雨幕、车灯和模糊城市建筑。站厅中段有三根粗大的混凝土立柱,柱脚带黄黑警示条,尽头是一段通往废弃站台的短楼梯。 冷白灯管是主要光源,其中两盏接触不良、偶尔闪烁;红色警示灯在墙面提供少量轮廓光。斜吹进来的雨水、潮湿薄雾、地面积水、立柱、护栏和闸机都要参与战斗。场景空间在所有镜头中保持一致,不得随切镜改变门、立柱、闸机和玻璃护栏的位置。 开场时主角位于站厅中段偏左,背后靠近检票闸机,面朝右前方;敌人位于右前方约三米处,背后是玻璃护栏。前半段战斗向站厅深处移动,中段绕过第二根立柱,后半段反向压回闸机区域。除非通过环绕镜头明确展示换位,否则敌人始终位于主角前方,保持清楚的行动轴线。 【动作原则】 动作由现代实战拳法、肘击、膝撞、低位踢击、近身控制、擒拿拆解和短距离摔法组成。主角冷静、精准、动作幅度小,善于格挡后立即反击;敌人力量更大、进攻性更强,擅长连续压迫和贴身撞击。敌人不是等待挨打的木桩,主角也不能全程无伤碾压。主角必须经历一次明显失势和一次危险化解,最后依靠判断和重心控制夺回主动。 每次进攻都必须呈现起点、运行路径、接触点和结果;格挡必须真正改变攻击方向;受击者的头部、肩膀、躯干、脚步和重心按照力量方向产生反馈。禁止隔空挥拳、拳脚穿透身体、没有接触却自行后退、无故旋转、连续空翻、悬浮和夸张飞行。动作速度快,但关键接触点必须清楚,不得用混乱残影掩盖动作。 【0—4秒:敌人抢攻】 第一帧已经处于战斗中,不进行对视、摆造型或绕圈试探。摄影机位于主角右后方约一米,接近肩部高度,采用轻微广角的第三人称ACT跟随视角。主角位于画面左侧偏中央,敌人从右前方快速逼近。前景是主角潮湿的肩膀和抬起防守的手臂,中景清楚呈现双方身体,背景是玻璃护栏、雨幕和城市灯光。 敌人右脚踩入水洼,水花向后飞散,同时右直拳攻击主角面部。主角小幅向左偏头,使拳锋擦过脸侧,左手从内侧拍开敌人手腕,右前臂立即架住敌人跟进的左摆拳。敌人不停顿,顺势用右肩和身体重量向前撞击。主角被迫后退半步,鞋底在湿地面短暂滑动,但立刻稳住。 敌人抬膝攻击主角腹部,主角双肘向内压住膝部,将冲击引向身体左侧,同时右脚切入敌人支撑腿外侧,为反击建立位置。摄影机跟随主角后退,敌人始终处于正前方。动作快速但不使用慢镜头。 【4—8秒:第一次反击】 敌人的膝部被压开,主角手臂从镜头前掠过形成遮挡,自然切换到两人右侧的平行移动机位。采用略低的中景,清楚展示脚步、髋部和拳肘路径。 主角左手控制敌人的右前臂,右拳从胸前短距离击中敌人右侧肋部。敌人的上半身因受力向左收缩,呼吸短暂中断,但立刻用左肘横击主角太阳穴。主角抬右臂挡住肘尖,前臂在碰撞中明显震动;随后转动肩膀,以左肘短击敌人胸口,再用右脚低位扫踢敌人左侧小腿。 敌人的左脚被扫得横移,但用右脚及时撑住,没有夸张腾空。他抓住主角上臂,将主角猛烈推向第一根混凝土立柱。摄影机随主角急速后退。主角后背接近立柱时用左掌撑住柱面卸力,沿柱边旋转躲避。敌人紧随而来的重拳擦过主角肩膀,砸中柱面,产生沉闷撞击、少量灰尘和疼痛反馈,但立柱不得碎裂。 敌人击中立柱时,镜头产生一次短促撞击震动。主角快速绕到立柱另一侧,摄影机紧贴主角背后绕柱跟随。立柱短暂遮满画面,完成隐形切镜。遮挡结束后,主角位于立柱右侧,敌人从左侧追出,两人沿站厅纵深向楼梯方向移动。 敌人连续使用左右直拳压迫主角。主角边后撤边格挡:右掌向外拨开第一拳,左前臂贴住第二拳。敌人第三次攻击改为右腿中段踢击。主角来不及完全避开,只能收紧左臂保护肋部。踢击真实命中护臂,力量让主角向右踉跄两步,肩膀撞上玻璃护栏。玻璃发出沉闷震颤却没有破碎,护栏外的雨夜灯光因撞击产生短暂拖影。 敌人跨步逼近,左手压住主角肩膀,右肘攻击面部。主角在最后一刻低头,肘部从头顶掠过;主角左手抓住敌人上臂,右手控制手腕,尝试将其压向玻璃护栏。敌人凭借力量强行抽臂,以额头短促撞击主角眉骨。主角头部向后偏移,额角出现轻微擦伤,目光短暂失焦。 【13—17秒:主角失势】 头撞形成明确节奏重音。快速切入主角侧面近景,表现接触瞬间、飞散雨水和真实眩晕,随即拉回能够看清全身动作的中景,不使用长时间慢动作。 敌人抓住主角上身,将他从玻璃护栏方向拽回站厅中央,右腿绊住主角左腿,试图完成向前摔投。主角身体倾斜,在倒地前用右脚重新踩稳,并以左手勾住敌人后颈,双方进入短暂贴身角力。 摄影机从主角侧后方环绕约九十度,明确展示两人换位:敌人转到画面左侧,主角转到画面右侧,主角背后变为第二根立柱和闸机方向。敌人以力量将主角压在立柱边缘,进行一次肩撞和一次短膝击。主角用大腿与前臂挡住主要冲击,但仍因疼痛弯腰。敌人抬起右肘,准备从上方向下砸击主角后背。镜头靠近主角肩后,使抬起的肘部在画面中形成强烈威胁。敌人的右肘快速下砸。主角不能凭空躲避,而是在肘击启动时向敌人身体内侧贴近,使肘尖从背后落空。主角左手托住敌人肘关节,右手控制其手腕,以肩膀作为支点向前旋转,破坏敌人的肩线。摄影机跟随主角完成半环绕,背景中的立柱、闸机和警示灯提供清楚的空间参照。 敌人为避免手臂被锁,顺势转身,以左拳反击。主角放开控制,低头闪过拳头,起身时用右肩撞击敌人胸口,将其推离立柱。主角立即以左直拳攻击敌人面部防守,迫使敌人抬高手臂;随后右拳从手臂下方击中腹部,再用左肘短击锁骨区域。三次攻击快速递进,但必须分别清楚,不能生成多条手臂。 敌人后退两步,脚后跟撞到检票闸机底座,随即用右腿横扫主角腰部。主角向前贴近,压缩踢击距离,以左臂承受大腿近端冲击,同时右手环住敌人腰侧,左脚踏到敌人支撑脚外侧。低机位镜头清楚交代主角如何控制敌人重心。 【22—26秒:环境摔投】 主角转动髋部并向斜前方发力,将失去支撑的敌人摔向闸机。敌人的肩背先撞在金属闸机侧面,闸机挡板被撞开并发出刺耳金属声,随后敌人落在湿地面,水花向外扩散。敌人的身体必须表现真实重量,不能轻飘飘飞行或在空中翻转多圈。 摄影机沿敌人落地方向快速俯冲,在触地时转为略高的斜俯视ACT视角。主角没有停下摆姿势,而是跨过被撞开的挡板继续追击。敌人翻身用右脚蹬向主角膝部,主角侧移避开。敌人借蹬腿动作翻滚起身,左手撑地,右脚踩稳,还未完全站直便挥出右摆拳。 主角从拳头内侧进入,左臂贴住敌人右臂,右掌推击敌人下颌。敌人向后仰头卸掉部分力量,同时左拳击中主角侧腹。主角明显收腹受击,却保持贴身距离,用右膝短促撞击敌人大腿前侧,进一步破坏其稳定性。两人呼吸加重,衣服逐渐湿污,但动作不能突然变得疲软。 【26—30秒:高潮反制】 摄影机回到主角左后方约一米的第三人称ACT跟随位置。主角位于画面中央偏左,敌人在正前方,背景保持被撞开的闸机、闪烁顶灯和红色警示灯。 敌人最后一次猛烈前冲,先用左拳虚晃主角面部,真正攻击是右肘横扫。主角没有被虚招骗出大幅动作,而是用右掌压下敌人左手,在右肘接近时向前踏入攻击内圈,左前臂架住敌人上臂,使肘尖从主角脑后掠过。主角抓住敌人后颈,右脚勾住其前脚踝,以肩膀向前下方施压。 敌人重心向后倒去,却抓住主角衣服试图将主角一同拖倒。主角立刻松开后颈,身体向左转出抓握,以右肘从极短距离击中敌人胸口偏肩位置,再用左掌推开敌人上胸。敌人向后撞在已经打开的闸机挡板上,挡板进一步弯曲。敌人半跪落地,一只手撑住湿地,另一只手仍保持防守,没有昏迷,也没有飞出画面。 主角受惯性影响向前半步,迅速重新稳住重心,双手保持防守,没有背对敌人、庆祝或摆胜利造型。摄影机缓慢移动到主角右侧,以主角肩膀作为前景,将焦点落在半跪喘息、暂时失去进攻能力的敌人身上。红色警示灯闪烁一次,远处传来雷声。最后一帧保持双方继续互相锁定的危险状态,让主角暂时占据优势,但为下一段战斗留下自然衔接。 【摄影、画面与声音】 保持电影级实拍质感与ACT游戏镜头的融合。主要采用肩后跟随、侧后跟随和腰部高度移动机位;关键命中使用短暂侧面中景或近景。摄影机移动快速但具有真实重量,不得漂浮、瞬移、疯狂旋转或持续无规则抖动。撞击时只使用极短、低幅度震动,震动方向与力量方向一致。 动作中允许适量运动模糊,但人物面部、手脚和接触点在关键时刻保持清晰。禁止鱼眼、长时间慢动作、子弹时间、冻结画面、速度线、游戏血条、准星、按钮提示和文字UI。ACT感来自主角中心构图、第三人称跟随、空间推进和及时的动作反馈。 超写实皮肤、肌肉牵拉、衣料褶皱、湿润头发、汗水和雨水反射;真实金属、混凝土、玻璃与湿地砖材质。冷蓝灰色为主色,冷白顶灯作为主光,红色警示灯作为局部轮廓光。高对比但保留暗部细节,高光不过曝,地面反射自然。禁止蜡像皮肤、塑料CG质感、噪点、脏污颗粒、锐化白边和压缩色块。 生成严格同步的现场声音:雨击顶棚、远处雷声、脚踩积水、湿地摩擦、衣料摆动、急促呼吸、拳掌击中皮肉、前臂格挡、肘膝碰撞、身体撞柱、玻璃震颤和闸机金属回响。无需对白。音乐仅保留极其克制的低频脉冲,不能盖过格斗音效。 【负面约束】 禁止身份互换、面孔漂移、服装变化、身体比例突变、多余人物、复制人物、多手多脚、关节反折、身体穿透和人物粘连;禁止隔空受击、动作没有惯性、敌人静止等待、无故滑行、瞬移、悬浮、武侠轻功、气功、能量特效、夸张冲击波和墙体爆炸;禁止主角全程无伤碾压,禁止敌人只会挨打;禁止无意义空镜、英雄登场、长时间对视、重复招式和结尾突然黑屏。 最终效果必须呈现清晰的动作结构:敌人强势抢攻、主角短暂反击、绕柱追击、主角失势、危险拆解、重新夺回主动、利用闸机完成摔投和最终反制。30秒全程高强度、无废秒,刺激而不混乱,凶狠但遵守人体力学,摄影机始终以主角为核心。

by lansenaiSource
Hover to preview · click for sound

Wuxia action comedy in a period inn

【风格】武侠电影级动作喜剧(Wuxia Action Comedy),古风客栈实景,8K,达芬奇高级调色,升格慢动作与快剪结合,真实布料与木屑物理,横屏16:9 【时长】30秒 【场景】古风江湖客栈大堂,木桌长凳、红灯笼、蒸笼冒热气,梁柱交错,光从窗棂斜射带浮尘 【角色】女侠A @图1(红衣冷艳,用剑)、女侠B @图2(青衣飒爽,用双筷/软鞭),两人武艺相当 [00:00-00:02] 钩子·同时出手 蒸笼掀开——热气散去,只剩孤零零一个白胖包子。两只手"唰"地同时探向它,在包子上方一寸僵住。镜头拉近两人对视,杀气骤起。 音效:蒸汽"嘶"声,一声兵刃出鞘。 [00:02-00:06] 拔剑·桌面开打 A拔剑挑翻整张桌子腾空,B踩着飞起的长凳跃上房梁,包子被气浪震到空中。碗筷茶壶四散飞溅。 [00:06-00:11] 空中争夺(升格) 包子在半空划弧,两人凌空交手,A剑尖挑、B筷子夹,包子在刀光间被抢来抢去,谁也没碰实,热气还在包子上冒。 [00:11-00:16] 客栈大乱 战场蔓延,两人踩桌翻梁、灯笼被削落、木屑纷飞,其他"食客"抱头躲桌下。包子始终在两人争夺中心悬着不落。 [00:16-00:21] 决胜一击(升格特写) A与B同时使出绝招,剑气与鞭影在包子两侧交汇——包子被劲气托着高高弹起,两人各自摆出接杀造型,屏息等它落下。 [00:21-00:26] 神反转 镜头跟着下落的包子——一只慵懒的橘猫"喵"地从桌上跃起,在半空一口叼住包子,稳稳落地,当着两人的面吧唧吧唧吃完,舔爪。 音效:所有打斗音效戛然而止,一声"喵~",乌鸦飞过。 [00:26-00:30] 傻眼收尾(定格) A和B保持着帅气的接杀姿势僵在原地,剑和筷子还举着,齐刷刷看向吃饱打嗝的猫,表情裂开。 字幕(砸出):「有些仗,打赢了也没用。」 音效:一记喜剧"咚",猫满足地打了个嗝。 【导演备注】1.反差铁律:打斗越是天崩地裂、招式越华丽正经,最后被猫截胡的落差越好笑;2.身份锁定:A=@图1红衣用剑、B=@图2青衣用筷/鞭,脸/服装全程一致;3.节奏:2秒对峙→拔剑→空中争夺→大乱→决胜→猫反转,动作段用快剪+关键处升格;4.物理真实:包子热气、木屑、灯笼、布料惯性、猫的重量都要真实,武侠不卡通;5.包子必须全程在画面中心被争夺,是视觉焦点和笑点核心;6.不见血、不伤人,纯招式对拆。 【负面提示词】卡通武侠,塑料CG,换脸,两人身份/武器混淆,包子中途消失,猫出现得太早,血腥伤人,多余手指,字幕提前,水印,真实品牌logo,画面糊乱,未成年人形象。

by johnAGI168Source
Hover to preview · click for sound

Duel at Hulao Pass, identities from reference images

Duration: 20s | 16:9 | 24fps Duel from Romance of the Three Kingdoms, battle at Hulao Pass, late Han dynasty. REFERENCES Image 1 — 呂布 LU BU Image 2 — 関羽 GUAN YU Image 3 — 張飛 ZHANG FEI Use them for IDENTITY ONLY: face, beard, hair, headwear, armour, cloak colour, build, proportions. IGNORE their white studio background, flat studio lighting and standing pose — none of it appears here. Every shot is outdoors, and all three men are mounted and fighting. STYLE (all beats) A dusty plain before a huge stone mountain pass at dusk. Two vast armies on either side, banners rippling, dust in the air. Overcast sky, low copper sun breaking through. Palette: iron grey, ochre dust, blood red, bronze. Desaturated and earthy. 50mm, shallow focus, heavy haze, film grain, high contrast, hard rim light through the dust. Photorealistic historical epic, sweat and grime, worn scratched armour. No fantasy glow, no magic. WEAPONS — no reference exists. Build from this, keep identical throughout. Each is 1.5x its owner's height. Never swap them. 呂布: ji halberd — straight spear point, one crescent side blade below it on one side, long dark shaft, bronze fittings. 関羽: guandao — one heavy broad curved blade with a hook on its back edge, joined to a long dark shaft by a bronze dragon head. 張飛: serpent spear — slender steel blade forged in an S-shaped wave, on a plain dark shaft bound with cord. HORSES — no reference exists. Anatomically correct, four legs, correct joints. 呂布: large red-chestnut stallion, black mane, bronze fittings, red tassels. 関羽: tall dark bay, black mane and lower legs, plain leather tack. 張飛: heavy solid black, thick neck, shaggy fetlocks, battered iron tack. BEAT 1 (0-4s) 呂布 — clean-shaven, gold headdress with two pheasant feathers, red cloak, halberd, red-chestnut horse — sits alone between the two armies, halberd across his saddle. The horse stamps and turns a half step. He raises the halberd and levels it at the enemy line. Camera: WS from low, slow push-in. BEAT 2 (4-8s) 張飛 — bristling beard, black iron armour, dark cloak, serpent spear, black horse — bursts from the line at a gallop, roaring, and drives the spear at 呂布. 呂布 turns the halberd across his body and knocks it aside. Sparks and dust. The horses wheel past each other. Camera: tracking alongside at horse height, moving with the charge. BEAT 3 (8-12s) 関羽 — long smooth black beard, ruddy face, dark green robe over armour, green headwrap, guandao, dark bay horse — rides in from the far side, rises in the stirrups and cuts down overhead. 呂布 catches it on the haft. Both strain, blades locked, horses shoulder to shoulder. Camera: MCU, orbiting once around the locked weapons. BEAT 4 (12-16s) All three circle in churning dust, weapons striking and turning each other aside, horses rearing. 呂布 in red and gold, 関羽 in green, 張飛 in black iron. Cut tight to 呂布's eyes under the headdress — narrowing, calculating, uncertain for the first time. Camera: handheld, close, drifting between them. BEAT 5 (16-20s) 呂布 wrenches his horse around, sweeps the halberd once to drive them back, and gallops for the pass, red cloak and feathers streaming. 関羽 and 張飛 pull up side by side, chests heaving, weapons still raised, watching him go against the low copper sun. Camera: WS, pulling out to reveal both armies. AUDIO War drums and massed shouting throughout, rising through beats 2-4. Hooves on dry earth, snorting horses, steel ringing on steel, creaking leather, wind and dust. A deep roar from 張飛 at 4s. Near-silence for a breath at 12s on 呂布's eyes, then the clash returns. No dialogue. AVOID white studio background, flat studio lighting, seamless backdrop, neutral standing poses, facial hair on 呂布, on-screen text, subtitles, watermarks, modern objects, fantasy glow, magic, wire-work floating, clean unworn costumes, faces/armour/weapons swapping between beats, weapons changing shape or length, horses with deformed or extra legs, extra or missing limbs, more than three warriors, mixed camera moves, jitter, flicker, ghosting.

by tanabe_fragmSource
Hover to preview · click for sound

Arena duel, written beat by timed beat

(0-4s) Wide shot of the duel arena — the musculer man and the old kung fu master face off, crowd murmuring in anticipation. Referee raises hand and starts counting down: eng::**"3... 2... 1..."**. (4-7s) Referee shouts "Go!" / gives the start signal. The kung fu master bows slightly. (7-9s) Right as the master bows, the musculer man swings his leg and kicks him straight in the head. The master collapses. The entire crowd gasps in shock, frozen in disbelief. (11-15s) The musculer man **English**: "I thought the fight started, why is this old man bowing?" @dreamina_ai #DreaminaCPP

by WolfRiccardoSource
Hover to preview · click for sound

Japanese game-show wipeout

A contestant balances across a spinning foam roller over a pool, wobbles, and wipes out into the water - a full Japanese game-show obstacle bit, crowd cheering in the stands. Comedy timing and a physical pratfall are genuinely tough for AI and this holds. Most people test a new model on hero shots - dramatic, slow controlled. A game-show wipeout is the opposite: fast physical comedy, a body losing balance, a crowd reacting water splashing on impact. That's a much harder thing to make land. Here's why a clip like this is a real test 1. Physical comedy depends on timing the failure. A pratfall only works if the wobble, the overbalance, and the fall happen in the right rhythm. Too smooth and it's not funny; too random and it's not believable 2. A reacting crowd has to sell it. Rows of spectators clapping and watching in the stands do the same job the crowd does in any live-event clip -the audience reaction is what makes it feel real 3. Water and impact are a physics stress test. Body hitting a surface, then a splash, is exactly the kind of contact-plus-fluid moment that used to fall apart instantly 4. It's a specific, recognizable format. Japanese game-show obstacle courses have a distinct look and energy -nailing the genre, not just a generic person falling, is the flex Why this matters • Comedy and physical failure are underrated benchmarks - everyone tests drama, far fewer test a convincing wipeout • The version leap is the real story: putting the Seedance 2.0 generation side by side is how you actually judge progress, not marketing claims • Physical comedy landing at all is a sign the model is handling timing and balance, not just pretty stills Worth being straight about • A side-by-side against the 2.0 version is the honest way to show a jump - a single clip alone proves less than the comparison does • A teaser clip is the best-case cut - the real test is running your own awkward, physical, timing-dependent scene through it Everyone benchmarks new models on the dramatic hero shot. The quieter test is whether it can make you laugh at someone falling off a foam roller -and that's a harder problem than it looks. Prompt 👇

by NexlowXSource
Hover to preview · click for sound

Climbing a flooding botanical tower in a night storm

Scene and Mood: A lone figure climbs upward through a flooding vertical botanical archive tower during a night storm, glass tiers shattering and cascading water below, forcing a desperate ascent toward the glass dome roof as lightning flickers outside. Frame Map: Shot 1, seconds 0 to 6, subject small in the lower third of a tall vertical tower interior, tiers of glass terrariums stacked upward through the frame, water sheeting down from above. Shot 2, seconds 6 to 12, subject center frame climbing a metal service ladder bolted to the tower's inner wall, glass tiers visible left and right, water pouring past. Shot 3, seconds 12 to 18, subject in the right third crossing a narrow catwalk as a glass panel shatters in the left third behind them. Shot 4, seconds 18 to 24, subject small in the lower third reaching the dome's base, storm sky and glowing drifting spores filling the top two thirds through the glass ceiling above. Shot 5, seconds 24 to 30, subject silhouetted center frame bursting through the dome into open night air, lightning lighting the wider tower silhouette below. Subject Lock: One adult figure, lean build, soaked dark hair plastered to the forehead, a heavy canvas work jacket dark with water over a plain thermal shirt, fingerless work gloves, mud and plant matter smeared on one forearm. Posture straining and climbing throughout, gaze fixed upward, hands gripping rungs, rails, and glass edges with visible strain. Same face, hair, jacket, and build locked across all five shots, no change at any cut. Cross Frame Rules: Only one figure occupies every frame across the runtime, nothing else enters or crosses. Screen position shifts per the Frame Map, but soaked jacket state and climbing posture carry unbroken across every cut. Movement: Shot 1, subject wades through knee deep water at the tower's base then grips a ladder rung and begins climbing at second 4, water sheeting down past the shoulders, loose leaves and broken glass drifting past in the current below. Shot 2, subject climbs rung over rung, jacket heavy and dragging, a terrarium panel cracks audibly at second 9 sending a spray of water sideways past the frame. Shot 3, subject sprints across the catwalk at second 14, a glass tier shatters behind them at second 16 releasing a wall of water and debris that just misses their trailing foot. Shot 4, subject reaches the dome's base at second 20, presses both palms flat against the glass overhead, glowing spores drifting slowly against the dark storm clouds beyond it. Shot 5, subject drives a shoulder through the weakened dome glass at second 26, bursting into open air as lightning flashes and rain drives sideways, glass fragments scattering outward around them. Last Frame: Subject holds silhouetted center frame at the dome's broken edge, chest heaving, one hand still braced on the shattered frame, storm clouds and a bright lightning flash filling the upper half, the flooded tower's tiered silhouette dropping away below into darkness. No on screen text, no captions, no signage typography, no rendered text in the frame. World Plate: A tall cylindrical botanical archive tower at night during a storm, tiers of glass terrariums stacked vertically and lit faintly from within by dim growth lamps, water sheeting down the tower's interior walls, a narrow metal ladder and catwalk system running up the center. A glass dome caps the tower, storm clouds and lightning visible through it, faint bioluminescent spores drifting in the air near the ceiling. Sound Bed: Diegetic only, rushing water and heavy dripping throughout, metal rungs clanging under gripped hands, glass cracking and shattering at intervals, distant thunder rolling, wind howling as the dome breaks, heavy strained breathing and grunts of effort, rain hammering the outer glass, a final gust of wind once the dome opens. Capture Realism: Depth via suspended water mist and falling spray between the figure and the tiered glass planes behind them. Moisture without shine, soaked jacket and hair read matte and heavy, not glossy. Per zone specular kill on skin, wet skin reads diffuse, no beauty light glow. Contrast curve, controlled highlights on the lightning and growth lamps, lifted but not crushed shadow on the figure, mid tones held cool and desaturated throughout. Camera Capture: Wide latitude cinema capture, vintage 2x anamorphic character on Shots 1 through 3 at 35mm, handheld with restrained breath, shifting to clean spherical character on Shots 4 and 5 at 24mm, wider and harder shake against wind, fine 35mm grain, desaturated rendition crushed toward deep green and steel blue, 24fps natural 180 degree shutter blur, 30s total across five hard cut shots.

by bmx_ai13Source
Hover to preview · click for sound

Three minutes of wuxia swordplay at a rain-soaked ruined temple

生成一段完整连续的3分钟写实古装武打电影片段。横屏16:9,2.35:1电影宽银幕构图,24fps,4K电影质感,冷峻低饱和蓝灰色调,真实物理动作,凌厉但克制的写实武侠风格。动作设计强调力量、速度、重心、惯性和空间利用,不使用仙术、轻功飞行或夸张能量特效。 故事发生在暴雨深夜的一座废弃山寺。青石庭院积满雨水,屋檐不断滴水,破损灯笼在风中摇晃,暖黄色火光与冷蓝色月光形成强烈对比。男主约三十岁,黑发高束,穿深靛蓝色窄袖短袍、黑色长裤和旧皮护腕,身形精悍,神情冷静,右手握一把带有磨损痕迹的窄刃长刀。二十余名蒙面敌人穿统一黑色夜行衣,从山门、回廊和屋顶逐渐包围男主,分别使用短刀、长枪、铁棍和锁链。所有人物外形、服装、兵器、伤势和位置前后保持一致,敌人数量随着战斗逐渐减少,不得凭空增加。 0-15秒:暴雨中的山寺全景,镜头穿过密集雨幕缓慢下降,越过残破屋檐和摇晃灯笼,落到青石庭院中央。男主背对镜头独自站立,雨水顺着刀鞘不断滴落。二十余名黑衣敌人从四面回廊、山门和屋顶无声出现,形成层层包围。镜头贴近男主右手,他用拇指缓缓顶开刀镡,刀刃露出一线寒光。敌人首领在回廊阴影中低声说道:“拿下。”男主抬起眼睛,平静回答:“让开。”伴随低沉雷声,第一批敌人突然冲出。 15-30秒:四名持短刀敌人从正面和两侧同时夹击。镜头采用贴身手持跟拍,男主侧身避开第一刀,用刀鞘撞开第二人的手腕,反手拔刀格挡第三人的横斩,金属碰撞迸出短促火星。他不做多余动作,连续使用肩撞、肘击、扫腿和刀背重击,将三名敌人迅速击倒。第四名敌人从背后扑来,男主听到脚步声,低头避开刀锋,转身用刀柄击中对方面部,再借对方身体挡住迎面刺来的长枪。镜头绕男主半圈,其他敌人立刻收紧包围,没有停下来等待。 30-45秒:六名长枪敌人组成半圆枪阵,同时向男主逼近,枪尖不断封锁他的退路。男主没有正面硬拼,而是快速退向庭院狭窄石阶,引诱枪阵收缩。第一排长枪同时刺出时,他踩着湿滑石阶侧身滑下,刀锋由下向上连续斩断两根枪杆,随后抓住断裂枪杆向侧面拨打,令另外三支长枪互相碰撞。他踩上一尊倾斜的石狮,借高度越过低扫的枪杆,落下时以膝撞击倒一名枪手。镜头从低角度跟随落地动作,雨水被双脚猛烈溅开。 45-60秒:屋顶上的敌人突然跃下,前后夹击男主。镜头迅速仰拍,两道黑影穿过灯笼和雨幕落向庭院。男主翻身滚过地面,刀锋贴着敌人的靴底掠过,顺势来到回廊入口。他利用廊柱阻挡多人围攻,让敌人无法同时展开。两名敌人一左一右贴近,男主用刀背压住左侧短刀,抬腿踢中右侧敌人的膝弯,再借廊柱反弹转身,将左侧敌人撞进后方人群。一个灯笼被刀锋割断,从空中落下,在积水中熄灭,画面短暂陷入半明半暗。 60-75秒:战斗进入第一次高潮。剩余敌人分成前后两队,在回廊中高速追击男主。镜头采用侧面长距离跟拍,男主一边后退一边格挡连续刀击,刀锋、雨水和火星在画面中交错。他突然踢翻长凳,迫使前排敌人跳跃躲避,自己从长凳下方滑过,起身后连续击中两人的腰腹和手腕。另一名敌人从柱后偷袭,男主用左手抓住对方持刀手臂,将其带向廊柱,迫使刀锋嵌入木柱,再用肩膀将其撞倒。动作紧凑、真实、清晰,不出现无意义旋转。 75-90秒:山门外的弩手突然射击。弩箭穿过雨幕击碎回廊木窗,男主迅速掀起一张木桌作为盾牌。镜头紧贴木桌另一侧,数支弩箭接连穿透桌面,箭头距离男主面部只有几厘米。男主顶着木桌向前冲撞,将两名敌人压倒在台阶上,随后把木桌推向弩手方向,趁遮挡翻滚进入大殿。敌人紧随其后冲进殿内。最后一支弩箭擦过男主左肩,划破衣袖,留下明显但不过度血腥的伤口。从此刻开始,男主左臂活动受到轻微影响,呼吸逐渐沉重。 90-105秒:大殿内部空间狭窄,残破佛像和木柱形成复杂障碍。一个身材魁梧的敌人挥舞铁棍迎面砸下,男主横刀抵挡,却被巨大力量震退,长刀脱手滑到佛像台阶下。魁梧敌人连续挥棍横扫,男主俯身闪避,铁棍击碎木柱表面。男主贴近对方,让长兵器无法发挥,用掌根、肘击和膝撞连续攻击对方胸腹,然后抓住铁棍中段借力旋转,将魁梧敌人摔过肩膀,砸碎旁边长桌。男主没有时间拾刀,只能夺下一根断裂枪杆作为短棍迎敌。 105-120秒:四名敌人在大殿中同时围攻。男主使用断枪进行近身棍法,挡住两把短刀,利用枪杆两端分别打击敌人的手腕、喉前和膝盖。镜头在木柱之间快速横移,通过柱子遮挡完成自然转场。男主故意露出破绽,引诱两名敌人同时进攻,最后一刻侧身闪开,让二人的兵器互相卡住。他用断枪横扫二人腿部,再踩住掉落的短刀,将短刀踢到手中。另一名敌人从佛像上方扑下,男主举起短刀格挡,巨大的冲击让两人同时摔出殿门,重新落入暴雨庭院。 120-135秒:男主在积水中翻滚起身,敌人首领终于从回廊阴影中走出。首领身材高瘦,使用一把带锁链的弯刃,动作冷静而危险。剩余敌人退到四周封锁出口。首领甩动锁链,弯刃贴着地面划过,激起一道水浪。男主连续后退闪避,弯刃缠住他的短刀并猛然拉走。首领第二次甩出锁链,缠住男主脚踝,将他拖倒在湿滑地面。男主在被拖动过程中抓住一块破碎瓦片,割开缠绕脚踝的布带,翻身避开从上方劈落的弯刃。弯刃重重砸进青石地面,火星和雨水同时炸开。 135-150秒:三名剩余敌人趁男主倒地同时冲来。男主已经疲惫,左肩受伤,动作明显不再轻松。他故意保持半跪姿势,引诱第一人靠近,突然抓住对方手腕,将其带入首领再次甩来的锁链轨迹。锁链缠住敌人的兵器,造成短暂混乱。男主利用湿滑地面低身旋转扫腿,使另外两人失去平衡,再用刀鞘和拳肘连续击倒。最后一名敌人试图从背后刺击,男主在积水倒影中看见对方,头也不回地向后撞肘,随后转身用掌根将其击倒。庭院里只剩男主与首领相对而立。 150-165秒:最终决斗。雷声骤然炸响,首领抽回锁链弯刃,发动连续快速攻击。镜头在广角全景和刀锋特写之间切换:弯刃横扫、锁链回旋、男主侧闪、格挡、翻滚,每一次动作都清晰连贯。男主捡回落在庭院边缘的长刀,用刀身缠住锁链,却被首领猛力拉近。两人在极近距离展开拳肘搏斗,首领击中男主伤肩,男主痛苦后退,长刀再次险些脱手。首领高举弯刃完成最后一击时,画面短暂升格,男主没有后退,而是突然向前贴身进入锁链内圈,用护腕架开首领手臂,以刀柄重击其胸口。 165-180秒:恢复正常速度。首领踉跄后退仍想挥刀,男主踩住落在积水中的锁链,旋身夺下弯刃,再用长刀精准架在首领颈侧,动作突然停止。雨声重新变得清晰,首领手中的兵器缓缓掉落,跪倒在地。男主没有补刀,只是收刀入鞘。镜头缓慢拉远,庭院中敌人全部失去战斗能力,兵器散落在雨水里,没有血腥特写。男主扶着受伤的左肩穿过山门,破晓的微光从远处云层后出现。最后一个镜头停留在积水中的刀痕和逐渐平静的波纹上,远处传来男主渐行渐远的脚步声,画面自然淡出。 镜头要求:整段战斗保持明确的空间方向和人物位置,使用大远景交代战场,中近景呈现攻防动作,适量使用贴身手持跟拍、低角度跟拍、环绕移动、遮挡转场和关键动作升格;不要频繁无意义快切,不要让镜头晃动到看不清动作。敌人必须采取包围、夹击、枪阵、远程掩护和多人协同,不能排队等待男主逐个攻击。男主通过地形、廊柱、石阶、长凳、木桌和敌人之间的位置关系化解人数劣势。 声音设计:全程保留真实暴雨声、雷声、急促脚步声、衣料摩擦声、呼吸声、刀剑碰撞声、木材碎裂声和兵器落地声。配乐以前半段低沉鼓点和弦乐持续积累,中段加入快速鼓点,最终决斗时音乐达到高潮;首领跪倒后音乐立刻停止,只保留雨声和脚步声。 禁止出现:仙术、飞行、能量波、武器发光、夸张爆炸、血腥肢解、现代物品、字幕、文字、LOGO和水印;禁止人物脸部突变、服装变化、额外肢体、兵器凭空出现或消失、敌人数量突然增加、同一个敌人重复出现、动作穿模、无重力漂浮、连续无意义空翻以及男主毫发无伤。整段必须像一场真正拍摄完成、叙事清楚、动作连贯、逐步升级的3分钟电影级武打戏。

by baqiceloudezhuSource
Hover to preview · click for sound

Thirty seconds of action in one unbroken take

A photorealistic live-action 30-second uninterrupted one-shot action sequence, presented in native 16:9 widescreen cinematic format. Designed specifically for a full-width IMAX composition with an epic sense of horizontal scale. Visual quality comparable to a $350 million Hollywood blockbuster. The camera is the true protagonist. A devastated futuristic Tokyo stretches across the horizon. Endless ruined skyscrapers, collapsed highways, burning buildings, thick smoke columns, drifting ash, military helicopters, and distant explosions fill the entire width of the frame. The opening begins with an enormous ultra-wide aerial establishing shot. The camera dives from the clouds toward the ruined city at breathtaking speed before seamlessly transitioning into an aggressive FPV drone flight. A lone samurai in realistic black armor stands in the center of an abandoned avenue. Far in the distance, occupying nearly the entire width of the horizon, a colossal 120-meter stone titan slowly emerges between skyscrapers. The camera never cuts. The titan charges. Every footstep sends shockwaves through multiple city blocks. The camera accelerates violently. It flies between skyscrapers, through broken office floors, underneath collapsing bridges, inches above destroyed streets, then suddenly climbs hundreds of meters before diving again. The camera constantly sweeps horizontally across the city, fully utilizing the 16:9 frame to reveal the immense scale of Tokyo. Buildings rush past both sides of the frame with massive cinematic parallax. The samurai sprints through the destruction. The camera races beside him, then overtakes him, flies backward while tracking his face, performs wide orbit shots, dramatic side tracking shots, sweeping crane movements, low-angle hero shots, extreme wide establishing shots, and breathtaking aerial reveals—all without a single cut. The samurai runs vertically along a collapsing skyscraper. The camera follows only inches behind him before swinging outward into an enormous wide shot, revealing the titan filling almost the entire frame. The camera dives beneath the titan, spirals upward around its massive body, then performs a huge 360-degree orbit as debris, fire, shattered glass, sparks, and smoke explode across the widescreen image. Natural cinematic camera inertia, realistic handheld momentum, subtle vibration from explosions, dynamic focus pulls, anamorphic lens flares, realistic motion blur, volumetric lighting, HDR, atmospheric haze, practical lighting, physically accurate destruction. Silence. The camera slowly pushes toward the samurai. He assumes an iaido stance. An explosive burst of movement. The camera follows the slash in one impossible continuous motion before rapidly pulling back thousands of meters. The titan freezes. A glowing cut slowly appears. The titan splits perfectly in half. Both halves collapse across the entire width of the city, generating an immense wall of dust that rolls across the skyline. The final shot is an ultra-wide aerial pullback revealing the burning ruins of Tokyo stretching from edge to edge of the frame, with the lone samurai reduced to a tiny silhouette in the foreground. Photorealistic live action, IMAX cinematic realism, ultra-detailed environments, massive sense of scale, seamless one-shot, award-winning visual effects, physically believable camera movement, spectacular destruction, emotionally powerful composition, 16:9 optimized, ultra-wide cinematic framing, no text, no subtitles, no watermark. A photorealistic live-action 30-second uninterrupted one-shot action sequence, presented in native 16:9 widescreen cinematic format. Designed specifically for a full-width IMAX composition with an epic sense of horizontal scale. Visual quality comparable to a $350 million Hollywood blockbuster. The camera is the true protagonist. A devastated futuristic Tokyo stretches across the horizon. Endless ruined skyscrapers, collapsed highways, burning buildings, thick smoke columns, drifting ash, military helicopters, and distant explosions fill the entire width of the frame. The opening begins with an enormous ultra-wide aerial establishing shot. The camera dives from the clouds toward the ruined city at breathtaking speed before seamlessly transitioning into an aggressive FPV drone flight. A lone samurai in realistic black armor stands in the center of an abandoned avenue. Far in the distance, occupying nearly the entire width of the horizon, a colossal 120-meter stone titan slowly emerges between skyscrapers. The camera never cuts. The titan charges. Every footstep sends shockwaves through multiple city blocks. The camera accelerates violently. It flies between skyscrapers, through broken office floors, underneath collapsing bridges, inches above destroyed streets, then suddenly climbs hundreds of meters before diving again. The camera constantly sweeps horizontally across the city, fully utilizing the 16:9 frame to reveal the immense scale of Tokyo. Buildings rush past both sides of the frame with massive cinematic parallax. The samurai sprints through the destruction. The camera races beside him, then overtakes him, flies backward while tracking his face, performs wide orbit shots, dramatic side tracking shots, sweeping crane movements, low-angle hero shots, extreme wide establishing shots, and breathtaking aerial reveals—all without a single cut. The samurai runs vertically along a collapsing skyscraper. The camera follows only inches behind him before swinging outward into an enormous wide shot, revealing the titan filling almost the entire frame. The camera dives beneath the titan, spirals upward around its massive body, then performs a huge 360-degree orbit as debris, fire, shattered glass, sparks, and smoke explode across the widescreen image. Natural cinematic camera inertia, realistic handheld momentum, subtle vibration from explosions, dynamic focus pulls, anamorphic lens flares, realistic motion blur, volumetric lighting, HDR, atmospheric haze, practical lighting, physically accurate destruction. Silence. The camera slowly pushes toward the samurai. He assumes an iaido stance. An explosive burst of movement. The camera follows the slash in one impossible continuous motion before rapidly pulling back thousands of meters. The titan freezes. A glowing cut slowly appears. The titan splits perfectly in half. Both halves collapse across the entire width of the city, generating an immense wall of dust that rolls across the skyline. The final shot is an ultra-wide aerial pullback revealing the burning ruins of Tokyo stretching from edge to edge of the frame, with the lone samurai reduced to a tiny silhouette in the foreground. Photorealistic live action, IMAX cinematic realism, ultra-detailed environments, massive sense of scale, seamless one-shot, award-winning visual effects, physically believable camera movement, spectacular destruction, emotionally powerful composition, 16:9 optimized, ultra-wide cinematic framing, no text, no subtitles, no watermark.

by tanabe_fragmSource
Hover to preview · click for sound

Aerial action spectacle, feature-film scale

30-second cinematic aerial action spectacle, breathtaking high-budget feature film, ultra-dynamic editing, fluid aerial choreography, premium cinematic visuals, seamless kinetic camera movement, aggressive speed ramps, rapid match cuts, whip pans, rotating camera moves, dramatic perspective shifts, cinematic lens distortion, volumetric storm clouds, ocean spray, glowing tracer lights, smoke trails, sparks, flying debris, colossal scale, emotionally epic, no dialogue. 0–5s Dark storm clouds gather above a vast open ocean. A colossal armored battleship cuts through enormous waves. Far above, a lone high-performance fighter aircraft emerges from the clouds. Extreme close-ups. Pilot visor. Hands on the controls. Jet intake vibrating. Radar display glowing. The camera alternates between immense aerial vistas and intimate cockpit shots. Complete silence. 5–10s The fighter rolls onto its wing and dives toward the ocean. The camera races only meters from the aircraft before instantly switching to first-person cockpit views, impossible wing-mounted perspectives, sweeping aerial fly-bys and dramatic low-angle shots skimming above the waves. The ship rapidly grows larger. The engines reach full power. 10–15s A breathtaking aerial montage. The fighter slices between towering smoke columns and dense clouds. Bright streams of tracer light streak across the sky. The aircraft performs impossible barrel rolls, high-G turns and vertical climbs while the camera continuously rotates around the fuselage. Speed ramps amplify every maneuver. 15–20s The fighter skims just above the ocean surface. Towering walls of spray rise behind it. It climbs sharply over the battleship's deck in one uninterrupted shot. The camera dives beneath the aircraft, then whips upward through drifting smoke, sparks and sunlight breaking through the storm. Everything feels impossibly fast. 20–25s The sequence reaches maximum intensity. The aircraft circles the massive vessel at incredible speed. Towering water plumes erupt around the ship. Clouds swirl overhead. The camera jumps between ultra-wide aerial compositions, intimate cockpit close-ups, spinning chase shots, dramatic slow-motion fly-bys and impossible orbiting movements around both aircraft and ship. Every cut increases the spectacle. 25–30s Everything suddenly slows. The fighter climbs vertically through the storm clouds into brilliant sunlight. Far below, the gigantic battleship disappears beneath drifting mist and shimmering ocean reflections. The camera pulls rapidly upward into an immense aerial shot revealing endless sky and sea stretching to the horizon. Hard cut to black. Sound design: soaring cinematic orchestra, thunderous percussion, deep brass, roaring jet engines, rushing wind, ocean waves crashing against steel, atmospheric echoes, powerful fly-by whooshes, subtle sonic booms, cinematic bass swells, brief silence before the final climb, overwhelming orchestral finale.40:T12c8,Use

by ArtedeingenioSource
Hover to preview · click for sound

WWI dogfight, camera specified shot by shot

【30 秒一战空战 严格分层镜头版提示词(逐镜头标注摄影参数 + 前景 / 中景 / 背景三层光影,防过拟合) 30 秒高镜头密度一战空中格斗完整故事线,激烈战场硝烟氛围,一战高空漫天混战场景,超大画幅宽银幕,粗颗粒复古胶片质感,黄褐色复古战场影调,21:9 宽银幕,画面全程动态流畅无卡顿,每一个镜头严格区分前景 / 中景 / 背景三层景深,前后景焦距完全不同,杜绝平面前后景同焦画面,彻底避免 AI 过拟合同质化效果。德国双翼战机  德国战机2 ,英国双翼战机  英国战机 ,德军飞行员  德飞行员 ,英国飞行员  英飞行员 。 0-2s:大远景高空俯视 焦距:14mm 光圈:f/5.6。前景:浓密高空硝烟缓慢流动擦过镜头下缘;中景:两架主角双翼战机(德方战机  德国战机2 在后,英方战机在前  英国战机 )一起在云缝里高速俯冲,后机死死咬住前机尾后;背景:数十架一战士战机群在远方天空各处缠斗,拉出密密麻麻白色航迹,厚重铅灰色积云铺满整个天空。自然光硬侧光,光线从侧上方斜切下来,硝烟边缘被阳光打透成半透明金色。 2-4s:驾驶舱中景跟拍 焦距:75mm 光圈:f/2.0。前景:螺旋桨转动拉出动态模糊气流从镜头前扫过;中景:德方飞行员  德飞行员 坐在驾驶舱里猛推操纵杆,头盔风镜反射出前方英方战机轮廓;背景:战机帆布蒙皮机身结构、高空远处模糊的云团轮廓层次虚化。侧逆光打亮飞行员肩线,驾驶舱内部保留暗部细节,浅景深背景完全虚化。 4-7s:机身侧面跟拍中近景 焦距:50mm 光圈:f/2.8。前景:高速气流卷动机翼蒙布线头抖动擦过画面边缘;中景:两架飞机机翼几乎贴在一起,德方战机机枪口连续喷吐出枪口焰,亮白色曳光弹线条连续擦过英方战机机翼蒙皮表面;背景:远处空中其他战机开火的微弱枪口焰光点零散分布在硝烟里。侧顶硬光,机枪开火瞬间枪口焰在机身蒙皮上投下跳动的光斑。 7-10s:机翼破口大特写 焦距:180mm 光圈:f/1.4。前景:飞溅的金属蒙皮碎屑在镜头前景慢动作划过;中景:英方战机机翼表面被机枪子弹击中,帆布蒙皮撕开破口漏出内部钢丝骨架,缕缕白烟从破口处冒出来;背景:远处蓝天和云层被浅景深完全虚化成柔和色块。侧光精准打亮破口边缘撕裂的帆布纤维纹理,暗部保留阴影层次,极致浅景深。 10-13s:飞行员近景特写 焦距:85mm 光圈:f/1.8。前景:高速气流吹动飞行服领口在画面左下角微微晃动;中景:英方飞行员猛拉杆回头看向后方,脸上紧绷全神贯注,快速扳动操纵杆做极限规避动作;背景:驾驶舱风窗外飞速流过的云层完全虚化。侧光半打亮飞行员半张脸,另一半落在硬朗阴影里,浅景深背景完全模糊,突出人物紧绷情绪。 13-17s:极限规避低角度环绕镜头 焦距:24mm 光圈:f/4。前景:地面炸起的橙色硝烟烟尘从画面底部向上翻涌;中景:英方战机侧身几乎 90 度侧翻,大角度转向从德方战机射击窗口里快速滑脱,机翼下方掠过地面一战战壕铁丝网轮廓;背景:远方地平线处数十个地面炸开的硝烟烟柱缓缓升起。侧逆光把战机机身勾出完整金色轮廓,低角度让机身在天空里极具冲击力,中景深只把机身主体拍实,远景硝烟层次自然虚化。 17-21s:高空拉升大远景 焦距:20mm 光圈:f/8。前景:半透明硝烟云团从镜头前缓缓飘过;中景:两架主角飞机一追(德方战机)一绕(英方战机)在高空云团里拉出交叉的白烟轨迹;背景:整片天空布满上百条战斗机拉出的白色航迹线,无数小黑点大小的战机在远景各处互相厮杀,庞大混乱的一战空战战场全貌完整铺开。顶光漫射下硝烟层次丰富,大景深完整收纳前景中景背景三层全部清晰的战场空间,画面纵深感拉满。 21-24s:迎头对冲低角度仰拍 焦距:35mm 光圈:f/2.8。前景:两架飞机高速旋转的螺旋桨扇叶边缘虚化形成动态扇面;中景:双机同时迎头对冲相向飞行,机身快速向镜头逼近,距离飞速缩短;背景:后方整片高空混战的机群轮廓全部虚化散焦。正仰拍逆光,机身被阳光完全勾勒出黑色剪影,只透出机翼轮廓的半透明帆布纹理,前景螺旋桨动态模糊增加速度感。 24-27s:隔舱对视大特写 焦距:100mm 光圈:f/1.4。前景:驾驶舱风窗外晃动的气流波纹虚化在镜头前;中景:两名飞行员隔着呼啸气流在驾驶舱里互相对视一瞬,时间微微放慢;背景:两侧飞速向后流走的天空和云团完全虚化成运动模糊拖影。侧软光只打亮两名飞行员半张脸,背景极致虚化,浅景深把所有注意力锁在人物眼神对视的瞬间。 27-29s:贴脸擦身极高速跟拍 焦距:40mm 光圈:f/3.5。前景:机翼尖擦过的气流扰动在画面边缘形成动态扭曲波纹;中景:两架战机几乎贴脸擦肩而过,机翼尖差几厘米就会相撞,气流吹动机翼表面蒙布剧烈抖动;背景:两侧向后飞速退去的混战战机全部拉出动态运动拖影。追焦跟拍全程只有两架主角战机机身完全锐利,前后运动物体全部带自然运动模糊,速度感拉满。 29-30s:两机反向散开收束大远景 焦距:21mm 光圈:f/5.6。前景:最后一缕高空硝烟缓缓从画面中间飘过;中景:两架反向飞远的双翼机背对着彼此向天空两个方向散开,身后留下两道漫长的白烟航迹,黄昏阳光把机身轮廓镀上一层金边;背景:远方天空仍在持续厮杀的整个战场,漫天航迹和硝烟构成庞大的一战空战史诗全景。黄昏暖橙色侧逆光斜照整个画面,烟的边缘是金的,云层底部是暖橙色,远处战场硝烟带着冷蓝灰调,冷暖对比宽银幕史诗感拉满。 全局约束:全程动作节奏紧凑激烈,一战战机细节完全写实,没有任何现代喷气式飞机,无多余字幕无多余文字,胶片颗粒均匀自然,所有镜头的光圈焦距参数对应生成正确景深分层,严格杜绝前后景同焦平面的扁平化画面,完全避免过拟合 AI 同质化效果。镜头要自然,不要生硬。】全片不要有音乐,不要有文字。4d:T1e24,type: anime_opening_continuous_camera_flythrough style: visual: dark fantasy anime, cel-shaded 2D, high contrast, full-color Japanese anime motion_quality: 24fps, 作画枚数多め, メリハリのあるスピードランプ tone: stylish, dramatic, kinetic, elegant, ominous direction: \u003e 日本のフルカラーアニメのオープニング。 高速なカット割りの代わりに、一筆書きのような連続カメラワークで、 複数のキャラクターを次々と捉えていく。 カメラは右へ左へ、上下へ、奥へ手前へと飛び回るローラーコースターPOVで、 一切止まらず、各キャラクターの最も映える瞬間を連続的につないでいく。 各キャラクターのポーズ・表情・背景・空間演出は、 そのキャラクターの性格、武器、属性、感情の核、物語上の役割から 最も魅力的に見える形をAIが自律的に判断して構成すること。 camera: motion: fast forward dolly with lateral drifts, rollercoaster POV, fluid aerial movement core_rule: \u003e no hard cuts, no static frames, no fades. 全編を通して一続きのワンカットとして成立させること。 speed_ramps: \u003e キャラクターのリビール直前に減速し、決めの瞬間で一瞬だけ視認性を確保し、 通過と同時に再加速する。 lens_feel: dynamic anime cinematography, dramatic perspective shifts, strong parallax final_exception: \u003e ラストのクライマックスキャラクター、または集合決めカットの直前のみ、 重みを出すために初めてカメラの運動を極端に緩める、 もしくは短く静止してもよい。 environment_logic: base_space: 無限に広がる異空間ヴォイド variation_rule: \u003e 各キャラクターのリビール区間ごとに、そのキャラクターの属性、感情、 象徴、戦闘スタイル、役割に最も合う環境要素をAIが選んで構成すること。 同じ空間を繰り返さず、毎回異なる空間的印象をつくること。 ただし全体としてはひとつの連続した異空間を飛行している感覚を保つこと。 palette_options: - deep crimson (#5A0000–#8B0000) + cyan mint accents + black shadows - midnight blue void + white petals + silver light - blood red mist + black feathers + gold rim light - violet smoke + shattered glass + cold cyan - 上記をベースに、各キャラクターに合わせて自由に再解釈してよい floating_elements_pool: - petals - feathers - blades - ribbons - chains - shattered glass - ash - sparks - torn paper - ink splashes - crystals - smoke - water droplets - leaves lighting: \u003e dramatic rim light, deep shadows。 光源設計はキャラクターの性格や役割に応じて変えること。 支配的なキャラには逆光や王者のような縁光、 儚いキャラには拡散光、 狂気のあるキャラには不安定な明滅光など、 感情の印象を光で補強すること。 input_characters: description: \u003e 以下にキャラクターのリストを与える。 人数は固定しない。2人でも10人以上でも対応可能。 各キャラクターについて、名前・性格・武器・属性・役割・感情の核・象徴モチーフなどを参照し、 そのキャラらしさが最大化する一瞬を構成すること。 format_example: - name: Character Name role: protagonist / rival / villain / mentor / key figure / support personality: "" weapon_or_power: "" attribute_or_theme: "" emotional_core: "" symbolism: "" priority: high / medium / low reveal_planning: global_rule: \u003e 入力された全キャラクターを、物語の高まりが自然に増していく順に並べ替えてよい。 基本的には、軽め・不穏・謎めいた導入から始まり、 中盤で躍動感と情報量を増し、 終盤で主人公・対立軸・ラスボス・最重要人物へ向けて熱量を上げること。 per_character_rule: \u003e 各キャラクターごとに以下を自動で設計すること: 1. pose: 静止ではなく「動きの最中の一瞬」 2. expression: 感情の核が伝わる表情 3. framing: 最も映えるカメラ距離・角度 4. background_variant: そのキャラに対応した空間演出 5. foreground_wipe: 次のキャラへつなぐ前景通過物 6. motion_relation: カメラに対して向かってくる / すれ違う / 見下ろす / 落下する / 回転する 等 uniqueness_rule: \u003e 各キャラクターで、体勢、シルエット、画面構図、視線方向、背景モチーフ、 前景の抜け、動きのベクトルを被らせないこと。 連続して似た演出を使わないこと。 duration_rule: \u003e 各キャラクターの見せ尺は均一でなくてよい。 priority が高いキャラクターほど、わずかに長く、印象的に見せること。 低優先キャラは短く鋭く見せ、高優先キャラは減速を強めて存在感を出すこと。 character_processing_loop: instruction: \u003e 入力されたキャラクター配列を上から順に機械的に処理するのではなく、 オープニング映像として最も気持ちよく流れる順番に再構成してよい。 それぞれについて以下を判断して映像化すること。 for_each_character: decide: - このキャラは支配的か受動的か - 攻撃 / 防御 / 逃走 / 佇み / 観察 / 詠唱 / 指揮 のどれが最も似合うか - 武器や能力をどう一瞬で象徴的に見せるか - 視線、手、指、髪、衣装、武器のどこを強調すべきか - カメラがそのキャラに向かうべきか、横切るべきか、追い抜くべきか - そのキャラに最も合う背景と前景モチーフは何か output: - pose - expression - framing - background_variant - foreground_wipe - transition_vector transitions: between_reveals: \u003e 花弁、羽根、刃、リボン、鎖、ガラス片、煙、灰、インクなどが 必ずレンズ手前を至近距離で横切り、それをワイプ代わりにして次のキャラへつなぐこと。 rule: \u003e 必ず「前景がレンズを横切る → 空間が変化する → 次のキャラが現れる」 という流れを守ること。 continuity: \u003e 前景ワイプは単なる切り替えではなく、 カメラが本当に次の空間へ飛び込んでいく感覚を生むように使うこと。 climax_logic: final_focus_rule: \u003e 最後の区間では、最重要キャラクター (主人公 / ラスボス / キーパーソン / 物語の中心人物)に到達すること。 そのキャラは他よりも重く、強く、象徴的に見せること。 ending_options: - 最重要キャラクター単体でカメラが初めて止まり、一枚絵として決まる - 主人公の先に最大の敵が現れ、緊張感のある静止に入る - 最重要キャラのあと、短く複数人の集合構図へ流れ込み、一枚絵で締める final_image_rule: \u003e ラストはポスターのように成立する構図にすること。 それまでの流動性と対比して、最後にだけ「止まる意味」があるように演出すること。 output_goal: \u003e 結果として、全体が一本の連続したアニメOPショットとして成立し、 各キャラクターが短時間でも明確に識別でき、 しかも一人ずつ違う魅力で刻み込まれること。 単なる紹介映像ではなく、「この作品世界の熱量と因縁」が伝わること。 negative: - hard cuts - static camera (except final climax hold) - fades - text - logos - subtitles - photorealism - 3DCG look - repetitive poses - repeated environments - flat staging4e:Ta6b,Full prompt: POV — Morning Commute in Coastal Japan | Ultra-Realistic, One Continuous Take First-person POV, no cuts, one unbroken handheld-feeling shot from our own eyes on the wa

by df_renoSource
Hover to preview · click for sound

Two martial artists clash, low-angle

Use @Image1 as the visual reference for the overall visual language: low-angle cinematic composition, two martial artists clashing in mid-air, warm sunlight, blue sky, airborne dust, dramatic fabric movement and realistic live-action kung-fu choreography. Create a 30-second, 16:9 cinematic kung-fu action sequence in one continuous long take. Exactly two original adult East Asian martial artists wearing loose off-white traditional kung-fu uniforms. The location is an empty old-fashioned Chinese restaurant before opening, with long wooden tables, stools, a narrow counter, bamboo baskets, stacked wooden crates, hanging paper lanterns, fabric curtains and large wooden shutters. Use the environment naturally during the fight. Tables, stools, counters and screens may be pushed, kicked, vaulted over or used as shields. The objects must feel heavy and physically believable. No tea, no teacups, no food close-ups, no weapons, no dialogue, no music. Use only impact sounds, footsteps, wood breaking, fabric snaps, wind and swirling dust. 0–3 seconds — HOOK: Start at floor level inside the restaurant. The front wooden door suddenly bursts open. One fighter slides across the floor underneath a long table toward the camera while the other fighter leaps over the counter behind him. A flying foot passes within centimeters of the lens. The opening must be instantly fast, dangerous and cinematic. 3–6 seconds: The sliding fighter rolls up and grabs the edge of the wooden table, flipping it upright as a shield. The opponent kicks the table aside and uses a stool as a step to launch into a high spinning kick. The camera tracks between the table legs and follows the movement at a low angle. 6–9 seconds: They exchange rapid close-range punches and elbow blocks beside the counter. One fighter slams the opponent onto the counter, but the opponent rolls across it, kicks a stack of bamboo baskets into the air and escapes through the falling baskets. 9–12 seconds: The fight moves through a narrow aisle between tables. One fighter jumps from tabletop to tabletop while the other sweeps the tables aside with powerful kicks. Wooden stools spin across the floor, curtains whip in the air and dust rises with every landing. 12–15 seconds: The opponent grabs a hanging lantern cord and swings around a wooden support beam, using the momentum to deliver a flying side kick. The other fighter blocks with a metal serving tray, creating a sharp metallic impact, then throws the tray aside and counters with a spinning back kick. 15–18 seconds — FALSE VICTORY: One fighter is driven through a folding paper screen and crashes into stacked crates. The standing fighter pauses in the dust, believing the fight is over. For one quiet second, only the swinging lantern and drifting dust remain. 18–21 seconds: The fallen fighter suddenly plants one hand on the floor and launches into a low spinning recovery. He sweeps the standing fighter’s legs, jumps over him and kicks him backward into the wooden counter. The counter shakes violently but does not completely break. 21–24 seconds: Both fighters rush toward each other through the wreckage. One uses the fallen tabletop as a springboard, while the other rebounds off the wall and twists through the air. Their bodies cross in a powerful mid-air collision directly in front of the camera. 24–28 seconds — BIG TECHNIQUE: The nearer fighter runs across the long wooden table, leaps onto the counter, grabs the hanging lantern cord and swings in a wide circular arc across the entire restaurant. At the highest point, he releases the cord and performs a spectacular 360-degree spinning double kick. The opponent blocks with the tabletop, but the force drives him backward through the front wooden shutters. Wood splinters outward as bright sunlight floods into the dark restaurant. 28–30 seconds — FINAL PAYOFF: The camera follows through the broken storefront. The defeated fighter rolls across the dusty street and stops. The victorious fighter lands in the ruined doorway. Dust slowly settles, the hanging shop bell swings and rings exactly once, and both fighters remain frozen in a tense final pose. End on the wide view of the destroyed restaurant, drifting dust and warm sunlight. Maintain consistent faces, bodies and clothing throughout. Exactly two fighters, no extra characters, no duplicated limbs, no distorted anatomy, no magical powers, no blood and no random destruction. The choreography must remain readable, practical and physically grounded, with the final swinging-rope double kick as the visual climax.51:Tcf0,1980s film, authentic

by renataro9Source
Hover to preview · click for sound

Contemporary dance meets spectral blade combat

High-end cinematic 3D realism fused with rapid contemporary dance and spectral blade combat. Explosive turns, deep backbends, floor slides, aerial contractions and sharp limb extensions flow directly into precise single-katana strikes, creating choreography that feels graceful, feral and impossibly fast. Dense shadow matter forms predatory silhouettes, lunging smoke trails and collapsing black shockwaves, while every blade arc tears crimson fabric-like energy through the air and disintegrates impacts into ink-dark particles. Aggressive orbiting camera movement, low-angle tracking, extreme foreshortening, sudden lateral acceleration, cold volumetric haze and wet reflective surfaces amplify the performance without sacrificing silhouette clarity. Realistic cloth dynamics, controlled motion blur, physical blade weight and feature-film rendering preserve elegance, danger and scale. The performance begins mid-attack as a shadow wolf leaps through the smoke and the dancer @[character ref] folds beneath it in a violent backbend, drawing her single katana through its body during the recovery. More wolves circle and attack from every direction as she accelerates through spins, floor sweeps, aerial kicks and whipping turns, each movement becoming a clean blade strike that shreds them into black vapor and crimson fragments. The camera spirals through the pack with her, ending in full motion as she launches into another rotating slash against the final wave.

by aimikodaSource

How to write great Seedance 2.5 prompts

  1. 1Write it in one fixed order. Subject, then action, then location, then camera work, then lighting, then picture texture. "A person walking" gets you a stock loop; the same subject with a camera move and a light direction attached gets you a shot. Creators who have spent real credits on 2.5 converge on this order.
  2. 2Thirty seconds is written as stages, not as longer sentences. The official guide splits a long clip into [STAGE ONE] / [STAGE TWO] / [STAGE THREE], each with an opening state, one main event, and an end state the next stage inherits. Padding a 10-second prompt with adjectives does not make a 30-second clip; splitting the event into inherited states does.
  3. 3One main action per time segment, described exactly once. Whether you use stages or plain 0-5s / 5-10s ranges, give each segment a single main movement. Two things go wrong otherwise: three actions crammed into five seconds all get compressed, and the guide is explicit that describing the same action twice degrades it.
  4. 4Give every reference asset a job and an exclusion. Not "here are my images" but "@image1 supplies the armour design and colour breakdown. Do not take the background." You can attach up to 50 assets; the more you attach without assigning roles, the faster stability drops.
  5. 5Name your subjects and reuse the name. Write <SOLDIER>, <SHARI>, <WOMAN> once and refer to that token every time after. Over 30 seconds, "she" and "the man" drift onto the wrong body — a named token is the most affordable identity lock there is.
  6. 6Write the sound as a cue sheet, not as a mood. Audio comes out of the same pass as the picture, so list what you want: ambience, the specific foley, dialogue with the language named, and what should be absent. Several prompts in this library end with a full sound list, and it is a large part of why they look finished.
  7. 7Ask for the imperfections you want. Handheld shake, autofocus that settles late, exposure breathing, framing that clips a face. Realism in this model is a list of specified flaws — leave them out and you get a clean commercial no matter how many times you write "candid".
  8. 8End with the negative list, and keep parameters out of it. Close with what must not appear: no subtitles, no watermark, no extra limbs, and — for reference-driven jobs — never render the reference sheet itself or duplicate the subject. Resolution, duration and aspect ratio are generation settings, not prompt text.

Prompt formula

Subject + action + location + camera work + lighting + picture texture + sound. Past about ten seconds, wrap that in timed stages: each stage gets one main action and an end state the next stage inherits, then close with the negative list.

Example: <BARISTA> pulls the last shot of the night in a closing café and sets the cup on the pass. Rain on the window, chairs already stacked, one warm pendant lamp over the machine and the rest of the room in shadow. The camera holds a low medium on the portafilter, then rises to a profile as the cup lands. Photoreal, warm tungsten against cold window light, fine grain. Sound: the grinder winding down, steam hiss, rain, one mug on wood, no music. Do not add customers, subtitles or a watermark.

Frequently asked questions

What is Seedance 2.5?

Seedance 2.5 is ByteDance's video generation model, unveiled on 23 June 2026 at the Volcano Engine FORCE conference. On the shipped API it generates 4 to 30 seconds in a single pass (or duration -1 to let the model pick the length), takes up to 50 multimodal reference assets in one job (up to 30 images, up to 10 video clips and 10 audio clips), adds video editing and video extension as task types so a product or background can be swapped without re-rolling the whole clip, and generates audio in the same latent space as the picture. Output is 480p or 720p — the native 4K at 10-bit colour announced at FORCE is not something the API exposes. ByteDance also shipped an official prompting guide for it, which we have walked through in English.

Are these prompts free to use?

Yes. Every prompt on this page is free to copy and run anywhere you have access to Seedance 2.5 — no login, no payment.

Who wrote these prompts?

Two kinds, and they are labelled differently. Prompts the creator published are kept in the language they wrote them in, credited, and linked to the original post — copyright stays with the author. For clips whose creator never published a prompt, we reverse-engineer one from the finished video and mark it "reconstructed": it describes the output, so it cannot recover negative constraints, exact dialogue or the reference-image workflow. It is a writing reference, not the creator's prompt. If you are an author and want an entry removed or a credit corrected, email us.

Where can I run Seedance 2.5 today?

On apimodels.app, as the model seedance-2.5, since 8 August 2026 — 480p $0.120/s, 720p $0.270/s, charged only on success. That is a change: probing our production Volcengine Ark account on 4 August 2026, every Seedance 2.5 model id still returned InvalidEndpointOrModel.NotFound, and we said so rather than sell a page that errors on generate. One key covers it alongside Seedance 2.0 (still the one to use if you need 1080p, from $0.092/s), MiniMax H3, Kling, VEO and Grok. The creators in this library ran these prompts in ByteDance's own Dreamina (Jimeng) and in Higgsfield.

How is writing for Seedance 2.5 different from Seedance 2.0?

The craft is the same; the structure is not. At 4-15 seconds you can write one paragraph for one action. At 30 seconds you have to split the clip into stages, give each stage one main action and an end state the next stage inherits, and lock identity with named subject tokens instead of pronouns — otherwise the model spreads the action evenly and identity drifts across the back half. The reference workflow also scales: 2.0 takes up to 9 images plus a reference video and audio, 2.5 takes up to 50 assets, which makes assigning each one an explicit job and exclusion much more important.

How long should a Seedance 2.5 prompt be?

As long as the clip needs and no longer. The prompts in this library run from about 1,700 to 4,600 characters, and the longest ones are the 30-second pieces — not because they are wordier, but because a 30-second clip has more states to pin down. If your prompt is long because it describes the same action three ways, cut it; the official guide is explicit that repeating an action degrades it.

Run these prompts on apimodels.app

Seedance 2.5 is live here alongside 160+ other image, video, audio and language models — one API, one key, pay as you go, with $0.10 of trial credit on sign-up.

This library is open source on GitHub

The prompts we reconstructed are published in full under MIT; author-written prompts are indexed with credit and a link to the original post. Open an issue to suggest more.

View on GitHub

Other prompt libraries