Upload a video or paste a Xiaohongshu / Douyin / Bilibili link — automatic shot detection and per-shot breakdown into structured prompts with dialogue, ready for Jimeng / Seedance.
These settings apply to the whole batch
Real output — full prompts you can copy straight into Jimeng / Seedance.
This 23-second video exceeds the 15-second limit, so it was split into two tasks — task 1 (0–12s, walls and trenches) and task 2 (12–23s, battle and explosion).
## 事件简述 镜头从城墙俯瞰战场，拟人化的兔将军在垛口后指挥；随后切到战壕，火光与烟尘中队伍冲锋，末段一次爆炸掀起尘土。 ## 场景图提示词 生成图片，4K高清，古风战场城墙，黄土夯筑垛口，远处烽烟，阴天冷光，9:16 ## 角色图提示词 生成图片，4K高清，拟人化兔将军，铁灰披风与皮甲，坚毅神情，半身像，9:16 ## 道具图提示词 生成图片，4K高清，木质攻城梯与铁盾，做旧质感，单体白底，9:16 ## 任务划分 本视频共 23 秒，超过 15 秒上限，划分为两个任务。 ## 任务1（12秒） 分镜1 — [城墙受困] 【00:00-04:00】，中景，平视，固定镜头 画外音（说话人1）沉声：「守住这道墙。」
Battlefield walls with animal characters; environment, characters and props each become image assets, and shots are split at the 15-second limit.
Upload a video, or paste a Douyin / Xiaohongshu / Bilibili link
Keep the defaults, or pick an output format (all-reference / first frame)
Get prompts, a dialogue script and per-shot analysis in 30 seconds — copy and go
Fast human motion is occasionally mis-detected; name the subject in a refinement round to fix it.
