Photo to Video AI: From Photos to a Finished Ad
A shoot list for four photos
| Photo | Show | Why it helps |
|---|---|---|
| 1. The hero | The thing itself, filling most of the frame, in focus | It is the subject: the ad and the end card are built around it. For a product, dish or garment it is the one redrawn as a clean shot |
| 2. Another angle | The same thing from the side or above | Extra photos become the places and details the shots are made in |
| 3. A detail | A label, a texture, steam, a stitch | A detail the shots can be made around |
| 4. In use or in place | The shelf, the table, the street it lives on | Becomes a location, so the ad happens somewhere |
Put the hero first. Up to four JPG or PNG files, each under 10 MB, of something you have the rights to. If the read cannot make out what is being advertised, it says so on the spot and nothing is charged: too blurry, too tiny, or too ambiguous are the usual reasons. Which photos work for an ad.
What the read does to each kind of subject
| Subject | What happens to the photo | Why |
|---|---|---|
| Product, dish, garment | Redrawn as a clean studio shot on white | So it can be lit the same way in every segment |
| Shop, restaurant, car, house | Kept as it is; any people in it are removed first | Cutting a place out of its surroundings destroys the signage, depth and scale; a model given a face copies it, so bystanders are removed |
| App or website screenshot | Used as it is, never redrawn; one screen only, no real face | Redrawing turns readable interface into mush |
Then a dialog says what the read understood: a category (Product, Food & drink, Apparel, Vehicle, Store / property, App, Website or Service) and a name field you can correct. If you typed one line about what it is and who it is for, the line wins when it disagrees with the photo: “our family restaurant” makes the photo a place, not a product. Nothing is charged until you choose how to make the ad. How to make an ad from one photo.
Pick the vibe for the subject
| Vibe | What it does | A fit worth trying |
|---|---|---|
| Epic | Scale, spectacle, a big reveal | A launch, a car, a place |
| Funny | A joke with the product as the punchline | Everyday products |
| Warm | People, feeling, a small moment | Family business, gifts |
| Minimal | The product, light, silence, one line | Skincare, jewellery, design |
| Twist | It looks like one thing, then it is another | Anything with a reveal |
| Mouth-watering close-ups | Food and drink up close: steam, pour, cut | Dishes, drinks |
| Show how it works | Use it, see the result | Apps, services, gadgets |
| A day in the life | Someone living with the product | Shops, clothes, a room |
The vibe also sets the look of the ad and the colours of its end card. It does not change the price; length and quality do. Vibe, length, format and quality.
Length: one, two or four segments
An ad is made in 15-second segments joined in the final cut with the end card at the end. 15 seconds is one segment and a single beat: the product and one idea. 30 is two. 60 is four, and is a small story. Every segment carries at least one spoken line, and the tagline is said out loud before the final freeze. Nothing legible is written into the picture, because video models cannot write readable text; the words that have to land are spoken or put on the end card. Format is landscape for sites and screens or portrait for feeds and stories, and it does not change the price. Quality is 480P, 720P or 1080P, with the all-in price of each shown in the menu.
What the script is not allowed to say, and the end card
Superlatives and absolute claims such as “the lowest price anywhere”, “number one in sales” or “cures everything” are stripped from the script before it is made, and a brand or product name that is itself a claim is softened. The ad ends on a card with your main image and up to four lines: Brand, Product name, Tagline and Call to action. Any line can be blank. The brand is copied letter for letter from a name printed on a sign or package, or from what you typed. To change a line later, use Edit end card and Assemble: the final cut is rebuilt without reshooting. The card is in the ad’s dialogue language, and the music is written for the ad and mixed under the dialogue, with nothing for you to pick. The end card · Where the music comes from.
One-click or expert, and what you pay
| One-click generate | Expert mode | |
|---|---|---|
| Who writes the script | AI | AI first draft; you edit or rewrite it |
| Who runs cast, sets and shots | AI | You, in the editor dramas use |
| Price | One all-in price, a cap | A smaller start price, then each step on its button |
| Time | About ten minutes | As long as you take |
| Script editable afterwards | No; the shots are | Until it is broken down |
The all-in price is the most the ad can cost: each stage freezes and charges its own actual price, so the total lands at or below it, and a stage that fails is refunded. Reading your photos is free. Ads never use the free first project. You cannot switch a one-click ad into expert mode or back; each is a fresh ad. One-click or expert mode · The one price is a cap.
After it is made
If one shot is wrong, open the finished ad in the editor, fix that shot and press Rebuild final cut; the other shots are not touched. A different length or format is a new ad. Cancelling works before shooting starts; credits spent on stages that began are not refunded. For the same photos in a story instead of an ad, see AI image to video. For the engine behind it, AI ad video generator.
FAQ
How do I turn photos into a video?
Open the Ad Video tab, add up to four photos of one subject with the best one first, add a line about what it is, and choose a vibe, a length and a quality. An ad of 15, 30 or 60 seconds is made.
How many photos can I use?
Up to four of the same subject, JPG or PNG, under 10 MB each.
Is reading my photos charged?
No. The read is free and tells you what it thinks the product is, so you can correct it before you pay for anything.
Will my product be redrawn?
A product, dish or garment is redrawn as a clean studio shot. A shop, car, house or app screenshot is kept as it is.
My shop photo has customers in it.
That is fine. People are removed and the place is kept.
Will my logo or price appear in the video?
Not reliably in the shots, because video models do not draw legible text. The brand, product name, tagline and call to action are on the end card, and the tagline is also spoken.
How much does a photo ad cost?
It depends on the length and quality you pick. The all-in price is shown on the button and is a cap, so the final charge is at most that.
Can I fix one shot afterwards?
Yes. Fix the shot in the editor and rebuild the final cut. The other shots are not touched.
Updated November 2, 2026