Make a product review video from one photo
Turn one product photo and one prompt into a selfie-style review video, with a picture of a person holding your product first, then a short video from that picture, then a new voice if hers sounds flat.
What you get
A selfie-style review of your product that nobody filmed: a person on a snowy summit at sunrise lifts your product to the camera, switches between the front and back cameras, and says a few short lines about it, with her own voice and the sound of the place.
- You start from one clear photo of your product and one prompt.
- First comes a picture of the person holding your product, which the video starts from.
- Then the video, made from that picture and the prompt.
- Then, if her voice sounds flat, a new voice on the same words. Her timing stays, so her lips still match.
The prompt
Vertical 9:16 selfie-style UGC review of [YOUR PRODUCT], filmed on an iPhone that switches between the front and back cameras, early sunrise light on a snowy mountain top, handheld and unpolished, like a friend showing off something new, real skin texture, no filters, no music, no captions. Setting: a snowy summit at sunrise above a sea of clouds, a light wind. A woman in her late twenties in a bright orange beanie and a dark green puffer jacket holds the [YOUR PRODUCT] from the reference image: [ITS COLOUR, MATERIAL, CAP OR LID, AND LABEL]. The [YOUR PRODUCT] keeps that exact shape, colour and lettering in every shot. What happens, and what she says: Front camera: she lifts the [YOUR PRODUCT] close to the lens, the clouds glowing behind her. "Okay, I filled this hours ago, and up here it's still steaming hot." Back camera, the phone propped on a rock, wide: she sits down and twists the lid off with both hands, and steam rises against the low sun. Close-up, back camera: the open [YOUR PRODUCT] in her hand, steam curling over the rim. "Look at that." Front camera again, the [YOUR PRODUCT] beside her cheek, wind in her hair: "Super light, doesn't leak. It's coming on every hike." The [YOUR BRAND] lettering stays sharp and legible and matches the reference image in every shot.
The prompt is a shot list: one line for the style, one paragraph for the place and the product, then each camera moment with her words in quotes, and a last line that holds the label still. Swap the place, the actions and her lines for what your product does and where it is used, and keep her lines short, a few words per moment.
Two ways to make it
- Press Start with your product photo. On the Image tab that opens, add your product photo with Add reference, pick a picture model in Model, and set Aspect ratio to the tall phone shape. Ask for the person from the prompt's setting paragraph holding your product close to the lens, then send it.
- On the Video tab, pick the newest Kling model in Model, and add the picture you just made as the Start Frame.
- Paste the prompt. Where it says "from the reference image", write "in the picture", because the video starts from your picture.
- Choose the Duration and the Resolution. The settings show how many spoken words fit the length you pick and what the video costs. Then send it.
- Watch it with the sound on. If her voice sounds flat, open the video, press Change voice and pick a voice from the list. Her words and her timing stay the same.
First connect Novoads to your AI. You do it once.
In Claude, press this button and approve it.
Add Novoads to ClaudeIn Claude Code, run this line in your terminal.
claude mcp add --transport http novoads https://novoads.ai/api/mcpIn Cursor, add this to ~/.cursor/mcp.json.
{"mcpServers":{"novoads":{"url":"https://novoads.ai/api/mcp"}}}In VS Code, add this to .vscode/mcp.json.
{"servers":{"novoads":{"type":"http","url":"https://novoads.ai/api/mcp"}}}In any other AI app, add a connector with this address.
https://novoads.ai/api/mcp
Here is a photo of my product. Make a selfie-style review video of it in Novoads: first a picture of a person holding it close to the camera on a snowy mountain top at sunrise, then a short video from that picture in which she switches between the front and back cameras and says a few short lines about it. If her voice sounds flat, change it to a warmer one.
Your AI uploads your photo, makes the picture, then the video from it, and can change her voice afterwards. Ask it to tell you what each step costs before it runs it.
Limits
- The video starts from a picture, not from the product photo alone, so the picture of the person holding your product comes first.
- Her lines have to fit the length you choose: the settings say how many spoken words fit.
- A new voice changes how she sounds, not what she says or how fast she says it.
- Look at the first frames at full size before you post. A label can smear while the product rises into the shot, and the first frame is the thumbnail people see before they press play. Cut those frames if it does.
- Check the label and the hands in every shot.