promptvv

promptvv reverse-engineers AI video prompts from existing videos. A user uploads a local file or pastes a public Douyin, Xiaohongshu or Bilibili share link, chooses a video type and an output format, then sets the target model duration limit. The output contains an event summary, image prompts for scenes, characters and props, and per-task shot lists with shot size, camera movement, duration, action, performance and dialogue that can be used in Jimeng, Seedance, Kling or Doubao. Longer clips are split into multiple tasks to fit single-generation duration limits, and results support conversational re-creation, batch submission, an API and MCP access.