Loading page...Загружаем страницу...正在加载页面...
AtlasRepo

Several pictures, one AI video: how we send @Image1…@Image7 to Kling (and six other models) in order

说明与用途

RefVid 的作曲器现在支持最多七张图片,自动标记为 @ImageN,并按模型翻译标签,实现一致的多图像视频生成。

问题

Animating multiple distinct images in a single AI‑generated video while preserving each image’s identity and context across shots.

结果

The composer now handles up to seven images per clip, automatically translates tags for each model, enforces per‑model limits before charging, and produces consistent multi‑shot videos.

工作流

  • 1. Drag or paste images into the composer; each receives a sequential @ImageN tag immediately.
  • 2. Compose shots, inserting tags in prompts (e.g., "@Image1").
  • 3. On render, the system orders images per model, computes final positions, and rewrites tags to the model’s required syntax.
  • 4. The API call includes image_urls and elements; the model generates the video.
  • 5. The studio displays the resulting clip and prompts for review.

证据

Every image you add becomes a tag, @Image1, @Image2, and a shot's text says which one it is about.

来源

The system rewrites every @ImageN token into that model's own name for that position.

来源

On the list models, when a clip is close to the cap, extra angles of a saved subject go first; a picture or a start frame never does.

来源

  • 记录日期: 2026-11-01T21:29:22.205Z