AI Image to Video Generator: The Settings That Matter
Every control, and what it does
| Control | Options | What it changes | Price? | After the ad is made |
|---|---|---|---|---|
| Photos | 1 to 4, the first is the hero | The subject, the picture on the end card, the places the shots are set in | No | A different set is a new ad |
| One line | Free text, optional | The direction of the story. It wins over the photo when they disagree | No | New ad |
| Product or place name | Shown for correction after the reading | The name used on the end card | No | Edit end card |
| Ad vibe | Epic, Funny, Warm, Minimal, Twist, Mouth-watering close-ups, Show how it works, A day in the life | How the story is told, the look of the ad and the colours of the end card | No | New ad |
| Length | 15, 30 or 60 seconds | How many 15-second segments there are: 1, 2 or 4 | Yes | New ad |
| Format | Landscape or Portrait | The shape of every shot and of the end card | No | New ad |
| Quality | 480P, 720P, 1080P | Sharpness of every shot | Yes, shown in the menu | Fixed for the ad |
| Mode | One-click generate or Expert mode | Who writes and runs the steps | A cap, or per step | Cannot be switched |
| End card lines | Brand, product name, tagline, call to action; any can be blank | The only readable text in the ad | A re-join, if changed later | Edit end card, then Assemble |
Notice what is on the right-hand side of the table. Almost everything is fixed when you press Make my ad, and a change of vibe, length or format means a new ad. That is why the reading and the settings are worth a minute: they are free to change until the button is pressed. Vibe, length, format and quality covers the four pickers on their own.
What the one line really does
The line under the pickers is optional, and it is the only control that is plain language. It is read together with the photo and wins when the two disagree: write “our family restaurant” under a picture of a dish and the reading treats the photo as a place, not a product. It also seeds the story. If you leave it blank, the chooser tells you the direction the AI picked instead, in a line that starts “No brief written”, so you can see what it will build before you spend anything.
A good line answers two questions in one breath: what it is, and who it is for. “A fragrance-free night serum for people with dry skin” gives the script a customer and a reason. “Great product” gives it nothing, and superlatives and absolute claims are cleaned out of the line, the tagline and the end card, so a line that leans on “the best” will not survive to the ad.
Length is segments, and segments are the unit of everything
An ad is made in 15-second segments joined in a final cut with the end card at the end. So 15 seconds is one segment, 30 is two and 60 is four. Each segment is shot on its own, which has three consequences you can use.
- Price follows segments. More segments means more shots to pay for, which is why length is a price control and format is not.
- A bad shot is a small problem. Open the finished ad in the editor, change that shot and rebuild the final cut; the other segments are not shot again.
- Length decides how much story fits. Fifteen seconds is the product and one idea. Sixty seconds has room for a cast and a small turn, and uses the places and people the reading found.
The tagline is spoken out loud before the final freeze, so the end card repeats what the viewer just heard. The end card.
Format and quality are the shape and the sharpness
Landscape suits sites, screens and pre-roll; portrait suits feeds and stories. The format is the shape of every shot and of the end card, and your product and place photos are used at their own shape either way. It never changes the price.
Quality is the only picker that spends money on pixels. 480P is for drafts and small screens, 720P is what most people ship, and 1080P is for big screens. The menu shows the all-in price beside each, so you see the step up before you take it. If the quality you picked is unavailable when shooting starts, the ad says so and offers to continue with one that is. The same logic runs through draft at low resolution, finish at high.
Three settings sheets for three real jobs
These are suggested settings, not rules. The behaviour in the first two rows of each column is how the reading treats that kind of photo; the rest is a sensible starting point you can change.
| A neighbourhood restaurant | A budgeting app | A night serum | |
|---|---|---|---|
| Photos | Shopfront first, then the signature dish, then the dining room | Two screenshots, one screen each, no faces | The bottle on a plain surface first, then a texture close-up and one in use |
| What the reading does | Category Store / property. Customers are removed, the place is kept | Category App. Used exactly as it is, never redrawn | Category Product. Redrawn as a clean studio shot |
| The one line | Our family restaurant, for the lunch crowd | What it does and who uses it | What it is and for whom |
| Ad vibe | Mouth-watering close-ups, or Warm | Show how it works | Minimal |
| Length | 15 seconds to start | 30 seconds: two beats, problem then screen | 15 seconds |
| Format | Portrait, for feeds | Landscape for a site, portrait for stories | Portrait |
| Quality | 720P | 720P | 720P |
| Check on the end card | Brand read from the sign, call to action such as “Book a table” | Product name and a call to action such as “Download the app” | Brand, product name, a short tagline |
For the app column, a screenshot must be roughly a phone or computer screen shape: a full-page capture is far taller than any video frame and is refused until you crop it. Which photos work for an ad.
Controls that are not there, and why
| You might look for | What happens instead |
|---|---|
| A music picker | None. The music is written for the ad, composed as one piece and mixed under the dialogue, then carried across the end card. To use your own, download the final cut and replace the track in an editor. Where the music comes from |
| On-screen text for the name and price | Video models do not write legible text, so ads are shot without it. The end card is where the name, tagline and call to action appear as readable words |
| A first-frame or last-frame picture | Not offered. Photos are references for the subject and places, and every shot is composed from the script |
| Your own footage | Not accepted by the ad route. Remake is the door for an uploaded video, and it reads the clip rather than reusing it. The spoken lines of an ad are written with the script and performed by the generated cast |
| A switch from one-click to expert mode | Not possible. Each is a fresh ad |
For what to do once the first cut is in, see AI image to video creator: finishing the ad.
FAQ
What can I control in an AI image to video generator?
Here: the photos and their order, one line about what it is, the ad vibe, length, format, quality, the end card lines and whether the AI or you runs the steps.
Which controls change the price?
Length and quality. The vibe, the format and the one line do not. The quality menu shows the all-in price beside each option.
How many photos can I use?
Up to four of the same subject. The first is the featured product and the picture on the end card.
Can I change the vibe after the ad is made?
No. A different vibe, length or format is a new ad. The end card lines and any single shot can be changed without starting over.
Can I choose the music?
No. Music is written for each ad and mixed in for you. To use your own, download the final cut and replace the audio in an editor.
What language is the ad in?
The chooser shows the dialogue and end card language before you press either button. Check it there; some languages note that voices may occasionally mispronounce a word.
Does the video start from my photo?
No. Photos are references. A product is redrawn as a clean studio shot, a place is kept and a screenshot is used as it is, and each shot is composed from the script.
How long does it take?
A one-click ad is typically ready in about ten minutes; longer ads take longer because they have more segments to shoot.
Set it once, read it before you press
Every setting is free to change until you press the button.
Updated November 2, 2026