You need a video for a listing or an ad. You do not have a studio, a camera operator, a model, or three days. This is an extremely common position, and it has more than one answer — the answers just differ in cost, quality and how obvious it is that you did not film anything.
Here are the four routes that actually work, honestly compared.
Route 1: Animate the photos you already have
The oldest trick and still the most underrated. You have product photos. Pan across them, push in slowly, cut on the beat, add text. The result is a video.
What it costs: nothing but time. CapCut, Canva and iMovie all do this free.
What it looks like: clean, a little static, obviously made from stills. Which is fine — half the ads on your feed are made this way and nobody minds.
When it works: listing videos, simple feature callouts, anything where the job is to show the product clearly rather than tell a story.
Where it falls apart: you cannot show the product being used. If the selling point is "this collapses to fit in a drawer", stills cannot carry it.
The one thing that matters: motion must be slow and deliberate. Fast zooms on a still photo read as cheap. A 6% push over four seconds reads as intentional.
Route 2: AI image-to-video
Feed a still into a video model and describe how it should move. The model generates frames that did not exist — the product sits in a scene, light shifts, the camera drifts.
What it costs: a few cents to a few dollars per clip depending on length and resolution.
What it looks like: genuinely like footage, when it works. Like a melting hallucination when it does not.
When it works: short clips, modest motion, the product held steady in frame. Four to six seconds of a subtle camera move is a high-percentage shot.
Where it falls apart: hands touching the product, text on packaging, and anything longer than about eight seconds. Ask for a hand to pick the product up and you will usually get six fingers.
The one thing that matters: describe camera motion, not product motion. "Slow dolly in, shallow depth of field" is a request a model can satisfy. "The woman opens the jar and applies the cream" is a request for trouble.
Route 3: AI-generated scenes from the product
Rather than animating your photo, the tool generates a new scene containing your product — on a bathroom shelf, in a kitchen, held by a person — and then animates that.
What it costs: similar to route 2, plus the image generation step.
What it looks like: the most like a real production of any option here, and the most variable.
When it works: lifestyle context. Showing the product where it belongs rather than on a white background is most of what a good product ad does, and this is the cheapest way to get it.
Where it falls apart: product fidelity. The model is generating your product, not copying it, so the logo can drift, the label can go soft, the colour can shift a shade. For an unbranded item this is invisible. For anything with packaging it is the whole problem.
The one thing that matters: generate the still first, check the packaging, and only pay for motion once the frame is right. The image is the cheap step. Iterating there instead of on video is the difference between an affordable workflow and an expensive one.
Route 4: Someone else's footage, licensed
Stock video of a category — skincare being applied, a kitchen being cooked in — cut with your own product stills.
What it costs: $20–80 per clip on the main stock sites, or free on Pexels and Mixkit with more searching.
What it looks like: polished and slightly generic. Viewers do not consciously recognise stock, but they often feel it.
When it works: B-roll and establishing shots around your real product footage. Context the viewer does not examine closely.
Where it falls apart: as the main content. If the whole ad is stock, you have made an advertisement for a category rather than for your product.
The one thing that matters: check the licence covers paid advertising. Many free licences do not, and ad platforms do occasionally care.
Which to pick
| Situation | Route | |---|---| | Listing video, need it today, zero budget | 1 — animate stills | | Product with heavy branding and small text | 1 or 2 — keep the real photo visible | | Need lifestyle context you cannot shoot | 3 — generated scenes | | Need fifteen variants to test | 3 — only route that scales | | Unbranded or generic product | 2 or 3 — fidelity risk is low | | Need a single flagship hero video | None of these. Hire a shoot. |
That last row is worth sitting with. None of these routes produces a hero asset that will carry a brand for two years. They produce serviceable, testable, replaceable video quickly. That is a different job, and for most sellers it is the more valuable one — but it is not the same job.
The parts nobody tells you about
Audio matters more than the video
A mediocre video with confident voiceover and clean music outperforms a beautiful video that is silent. Most feeds autoplay muted, so add captions too — but the people who do turn sound on are the people closest to buying.
Vertical from the start, not cropped afterwards
A 16:9 video cropped to 9:16 loses a third of the frame and the composition goes with it. Decide the aspect ratio before you generate anything.
The first second decides everything else
Whatever route you pick, the opening frame determines whether anyone sees frame two. A clean product shot with motion already underway beats a slow fade from black every time.
Four seconds is usually enough
The instinct is to make it longer. Resist it. A four-second clip that lands beats a twelve-second clip that explains.
A practical sequence
If you want a workable video today, and you have one product photo:
- Write the claim first. One sentence. "Fits in a carry-on." If you cannot write it, no amount of video will rescue the ad.
- Pick the route from the table above based on whether your product is branded and whether you need context.
- Generate or assemble the still frames before touching video. Stills are cheap to redo.
- Add four to six seconds of motion to the frame that best supports the claim.
- Add voiceover and captions. The claim, spoken, in the first two seconds.
- Watch it on a phone, muted. If the claim is not obvious without sound, fix the captions.
- Make two more versions with different openings. Which one wins is not predictable from the armchair.
Where tools fit
Routes 2 and 3 need a generative tool. Faysell does route 3 specifically for e-commerce: a product photo or listing URL goes in, the product gets analysed, a script gets written around a selling point, and the shots are generated with the product reference carried through all of them.
The image step is deliberately separated from the video step, for the reason given above — iterating on a still costs a fraction of iterating on a clip. If you want to try route 2 directly, the Playground takes any still and animates it without the ad workflow.
If your file ends up too large for wherever you are uploading it, the free video compressor handles that in the browser, no account needed.
The summary
You do not need a camera to make a product video, and you have not needed one for a couple of years now. What you need is a clear claim, the right route for your product type, and the discipline to iterate on the cheap step before paying for the expensive one.
Start with the stills. Always start with the stills.