How to Write a Shot List for AI Video (With a Template)
A shot list for AI video is a table that breaks your script into single-action clips, one row per shot. Each row records the shot number, the script line it covers, the subject and action, the camera framing, the setting, the planned length and the prompt. Plan clips of a few seconds each, keep one action per shot, and write the list before you generate anything.
Why AI video needs a shot list
Most AI video tools generate short clips, not whole scenes. A finished video is therefore a sequence of separate generations that you cut together. Without a plan, you tend to generate whatever looks interesting, then discover you are missing the shot that connects two ideas.
A shot list fixes this. It also saves credits and time, because you generate only what the edit needs, and it keeps your visual style consistent from the first clip to the last.
The columns to include
| Column | What to write |
|---|---|
| Shot number | A simple running number so the edit order is obvious. |
| Script line | The exact narration or on-screen line this shot covers. |
| Subject and action | Who or what is on screen and the single thing they do. |
| Framing | Wide, medium, close-up, over-the-shoulder, aerial and so on. |
| Camera movement | Static, slow push-in, pan, tracking. Say "static" if nothing moves. |
| Setting and light | Location, time of day and light quality. |
| Length | Planned seconds on screen. |
| Prompt and status | The final prompt, plus a note such as "approved" or "regenerate". |
How long should each shot be?
Plan each shot to be only as long as it needs to be to carry one idea. Many tools produce clips of a few seconds, and quality often drops as clips get longer, so short shots are both easier to generate and better for pacing. Check the clip length your tool actually supports and plan to that limit.
A useful check: read your narration aloud with a timer. Each sentence usually maps to one to three shots. If one sentence needs one long shot, consider covering it with two shorter angles instead.
One action per shot
The most common shot-list mistake is packing a sequence into one row, such as "she walks in, picks up the phone and smiles". Models handle one clear action far more reliably. Split it into three rows: walking in, picking up the phone, the smile in close-up. You can cut between them in the edit.
A copy-ready example
Here is a short example for a 15-second video about a morning routine. The content is illustrative, so adapt it to your own story.
| # | Script line | Subject and action | Framing | Length |
|---|---|---|---|---|
| 1 | Mornings used to be chaos. | Cluttered kitchen counter, a hand grabbing keys | Close-up, static | 3s |
| 2 | Then I planned the night before. | Person writing a list at a desk lamp | Medium, slow push-in | 4s |
| 3 | Now it takes ten minutes. | Person leaving the house, door closing | Wide, static | 4s |
| 4 | Try it tonight. | Notebook with a checked list | Close-up, static | 4s |
Add a style line to every prompt
Write one fixed style sentence, such as the colour palette, lens feel and lighting mood, and paste it into every prompt. This is the simplest way to make separate clips look like they belong together. For the structure of each prompt, use the 7-part prompt formula, and for character matching see how to keep characters consistent.
Frequently asked questions
What is a shot list in video production?
A shot list is a table of every shot a video needs, usually with the framing, action, setting and planned length for each. It turns a script into a checklist you can shoot or generate.
How many shots does a short AI video need?
It depends on the length and pacing. A 15 to 30 second video often uses somewhere around five to ten shots, but let the script decide rather than aiming for a number.
Should I write the shot list before or after the script?
After. Finish the script first, then build the shot list line by line so every shot has a reason to exist.
Can AI write my shot list?
A chat assistant can draft one from your script. Review it carefully, because it may suggest actions or camera moves your video tool cannot produce reliably.