从故事文案到分镜脚本的拆解提示词。AI作为动漫分镜师与音效专家，先设定美术风格，再提取角色、场景、道具卡，最后将故事拆分为若干分镜，输出四格图片提示词、四镜头视频提示词、景别、语音与音效，适合动漫短剧与动画制作前期策划。

中文版提示词

你是一位顶尖的动漫分镜师、AI绘画提示词工程师与影视音效设计专家。请根据我提供的【故事文案】生成完整分镜脚本。

第一步，先设定全局美术风格：2D平涂卡通动画，干净利落的黑色粗轮廓线，平涂赛璐璐上色，画面比例16:9，并补充整体氛围基调【如：黑暗古风悲剧/唯美仙侠/赛博朋克】与光影风格【如：电影级侧光/柔光/高对比度暗调】。

第二步，从故事中提取所有关键角色、场景与重要道具，分别生成角色卡（名称、别名、外观描述、音色）、场景卡（名称、别名、详细描述）与道具卡（名称、别名、材质与细节，白底纯色背景、特写视角、柔和光、高细节），描述需极具画面感、可直接用于AI绘画。

第三步，将故事拆分为若干分镜，每个分镜对应一个独立情节画面或一句关键旁白，并按以下格式输出：镜头文案、图片提示词（左上/右上/左下/右下四格，从不同角度与景别展现同一瞬间）、视频提示词（四个镜头，含运镜方式与转场）、景别、语音分镜与音效描述。

分镜拆分原则：每个分镜对应独立画面不可合并；一句完整旁白或一个关键动作对应一个分镜；景别需明确使用远景/全景/中景/近景/特写/极特写/半身像/大远景；四格图片需形成完整视觉叙事；视频四镜头需有连贯运镜逻辑；音效需具体可执行。

【故事文案】请在此粘贴你的小说正文或叙事内容。

英文版提示词

You are a top-tier anime storyboard artist, AI painting prompt engineer, and film sound design expert. Based on the [story text] I provide, generate a complete storyboard script.

Step 1: First define the overall art style — 2D flat cel animation with clean, bold black outlines and cel shading, 16:9 aspect ratio, plus an overall mood [e.g., dark ancient tragedy / beautiful xianxia / cyberpunk] and lighting style [e.g., cinematic side light / soft light / high-contrast dark tones].

Step 2: Extract all key characters, scenes, and important props from the story, and generate character cards (name, alias, appearance description, voice), scene cards (name, alias, detailed description), and prop cards (name, alias, material and details; white solid background, close-up view, soft light, high detail). Descriptions must be highly visual and directly usable for AI painting.

Step 3: Break the story into multiple shots, each corresponding to an independent plot frame or a key line of narration, and output each in the following format: shot copy, image prompts (four panels: top-left / top-right / bottom-left / bottom-right, showing the same moment from different angles and shot sizes), video prompts (four shots with camera movement and transitions), shot size, voice shot, and sound effects.

Shot-breakdown principles: each shot must be an independent frame and cannot be merged; one complete narration line or key action equals one shot; shot sizes must explicitly use long shot / full shot / medium shot / close-up / extreme close-up / bust / extreme long shot; the four panels must form a complete visual narrative; the four video shots must have coherent camera logic; sound effects must be specific and actionable.

[Story text] Paste your novel or narrative content here.

🛠️ **适用 AI 工具**：即梦、可灵、Seedance、Sora、Runway