Guide
AI images or video clips: what should you use?
Choose motion when the action is the message. A still image with narration may be enough for an explanation, while a demonstration can need footage that shows the movement accurately. Match each shot to its job before choosing a model.
By LoopBatch · Checked 7 October 2026
- What you will have
- A visual plan that fits the story and its budget.
- What you need
- A script or shot list and any footage you already own.
- How long it takes
- One pass through the shot list before generating visuals.
Before you start
- Identify the shots where real movement is essential to understanding the subject.
- Check rights and permissions for any existing footage, people or product images.
Choose by what the shot needs to show
| Shot needs | Starting choice | Review before publishing |
|---|---|---|
| A narrated idea or diagram | Generated or original still | Accurate labels and readable detail |
| A physical process or product in use | Your recorded footage | The action is shown correctly |
| Atmosphere or an imagined scene in motion | Generated clip | Consistency, artifacts and AI disclosure |
| A real person or real event | Verified source material | Permission and factual accuracy |
Compare the generation components
For six five-second scenes, the current standard-image example is 21.6 credits and the generated-video example is 576 credits. This compares the named generation examples, not every image model with every video model.
An edited still can pan or zoom without becoming a new AI video generation. That editing or rendering still belongs in the full workflow estimate. Existing footage has no new AI visual-generation charge in this planning model.
Generation examples checked 6 October 2026. Rates include the published markup. One credit represents one US cent of metered spend. Dollar equivalents describe credit value, not a subscription invoice. Writing, narration, rendering, review, unused generations and retries need their own allowance.
Build a mixed shot list
Imagine a thirty-second explanation with five still illustrations and one five-second motion scene. Its visual-generation component is 114 credits at these example rates, before other steps. This is an illustrative edit, not a saved preset in the calculator.
The calculator applies one source to the whole scenario. For a mixed edit, calculate each group of generations separately and add them. Keep clips you can actually use rather than budgeting only for the final screen time.
The steps
Mark each shot as existing footage, a still or a motion generation.
Count the generations in each group and compare their credit components.
Add the other workflow steps and any unused attempts to the budget.
Review motion, factual detail and disclosure before approving the final video.
Check the result
- The essential action is visible and plausible.
- Generated words, faces and product details have been inspected.
- The forecast includes the cost of alternatives you expect to discard.
If something goes wrong
- If generated movement changes the subject incorrectly, use verified footage or redesign the shot as a still explanation.
- If a still feels slow, shorten it or change the edit before buying another generation.
Sources and assumptions
These sources were checked on the dates shown. Product limits and planning advice are identified separately from platform requirements.
- LoopBatch generation examples and credit assumptions, read 6 October 2026
- LoopBatch calculator: five-second scene assumption, read 7 October 2026
Common questions
Can I mix footage and generated visuals?
Yes as an editing plan. Budget each generation group separately and confirm what your chosen workflow supports before running it.
Related pages
Put the plan into practice
Build a workflow you can run again.
Describe your content, review the estimate and approve the finished post before it goes out.