A scene plan is a list of every picture in the video, in order, with the word each one starts on. It is the cheapest step in making a video and the one that saves the most. Every picture you make without a plan is a guess, and AI pictures cost money every time you guess again.
1. Split the script into sections
Mark where each part of the script begins: the hook, each main point, the example, the ending. Each section gets its own set of pictures. If a section is longer than a minute, it probably has two ideas in it.
2. Mark where the picture changes
Read each section and mark the word where the idea changes. That is where the picture changes. Not every five seconds, and not on every sentence: on the idea. A good rhythm for narrated video is a new picture every few seconds while the story moves, and a longer hold when the viewer needs time to read or think.
3. Describe each picture in one sentence
For every scene, write what the viewer sees: who, where, doing what, from how far away. Write it so someone else could make the picture from your sentence alone. Add any words that appear on screen, such as a number or a short label, separately from the picture itself.
4. Mark what repeats
Go through the plan once more and look for scenes that show the same place, person or moment as an earlier one. Mark them "same as scene 4", for example. Those are pictures you never pay for twice, and they keep the video consistent.
From our own work: in one of our early videos, the plan asked for thirteen pictures of what was really the same scene. Marking repeats would have made one.
5. Count and price it before you make anything
Count the new pictures the plan needs, after repeats. Multiply by what one picture costs you. That is the price of the video's pictures, and you know it before spending anything. If it is too high, merge scenes or reuse more, on paper, where changes are free.
6. Check the plan against the script
Read the script and the plan side by side. Does every picture match what the voice is saying at that moment? Is there any stretch with no new picture for too long, or three changes in two seconds? Is any fact shown in a picture that the script doesn't say? Fix it now, while it costs nothing.
Where vbe | video build engine fits
vbe is a Mac app that plans every scene from your script this way: the word each scene starts on, what it shows, the words on screen, and which pictures repeat. It counts the new pictures and shows their cost before any are made. See how it works.
Written by Aman, founder of vbe | video build engine and maker of rethink normal. Published 24 October 2026.