Write an AI-assisted video script and shot list

Write who the video is for, what it should explain and what the viewer should do next. A logline should describe the main change or argument rather than a list of attractive shots.

Set the delivery format and approximate length. Keep factual claims, names and required information in a separate source list so that a generated script cannot silently invent them.

Divide the story into an opening, a development and a conclusion. Each beat should have a visual purpose or a line of narration that advances the idea.

Ask a text model for alternative structures or a shorter version of an existing passage. Review the result against your brief; fluent wording is not proof of factual accuracy.

For each shot, note the subject, action, framing, source files and required output. Separate exact captions, diagrams and title cards from generative footage.

A narration paragraph can span several simple shots. Conversely, an important visual action may need time without more speech. Use a rough edit to test the relationship.

• What must be visible?
• Which input establishes the subject?
• What is generated, and what is edited?
• How will the shot connect to the next one?

Read the narration aloud and inspect the storyboard. Remove repeated ideas and shots that do not help the audience. Fix uncertain claims before they become part of a voiceover or graphic.

Keep the approved script and sources with the project. Generate the next shot from a bounded brief, then revise the sequence using what actually works rather than forcing every first draft into the edit.