wiki / video-multi-shot-continuity-sequence / edit

Suggest an edit.

Improvements are reviewed before going live, like any wiki.

Title

One-line summary

The prompt

Plan a [NUMBER, 3-6]-shot AI-generated video sequence about [SUBJECT AND STORY BEAT, E.G. A COURIER DELIVERING A PACKAGE ACROSS A RAINY CITY AT NIGHT], total runtime [SECONDS] seconds. Video generators produce one clip per prompt with no memory between clips, so continuity must be carried entirely by the prompt text.

First, write a continuity block: a fixed description of 40-70 words covering the recurring character (face, age, build, hair, wardrobe with colors and materials), the recurring environment, the lighting regime (source, direction, color temperature), the color palette, and one lens or style reference. Word it as literal visual description, not narrative, because it will be pasted verbatim into every shot prompt.

Then write one self-contained generation prompt per shot. Each must include: the continuity block verbatim; a shot size and camera movement unique to that shot; the specific action, written so it begins and ends within [MAX CLIP LENGTH, E.G. 8] seconds; where the subject enters and exits frame, chosen so consecutive shots cut together (exit frame left, enter frame right, or a match on action); and the sound for that shot (one ambient bed plus at most one effect).

Constraints: vary shot size across the sequence with at least one wide, one medium, and one close-up; keep screen direction consistent; never require two clips to show the same background detail from different angles, because generated backgrounds will not match — cut to a different angle or a close-up instead.

Finish with an edit note for each cut: the cut type and what motivates it.

Write it so someone can paste it straight into Claude, ChatGPT or Gemini and fill the [BRACKETS].