A 15-second AI promo video, shot by shot

A 15-second vertical video with captions and sound that you can post, plus a repeatable shot plan.

Best for: Small brands and creators who want a polished short promo without a film crew.

• One message: what should a viewer remember or do after 15 seconds?
• Any real product photos or footage you want to include.
• A video model that accepts a starting image, and a regular editor.

1. 1

   Write five shots, three seconds each

   A hook, a problem or desire, the product in use, a detail or benefit, and a call to action. Write each as one sentence you could draw. Short plans produce better AI clips than abstract ideas.

2. 2

   Create a still for every shot first

   Generate or photograph the opening frame of each shot. Fixing a still is cheap; fixing a rendered clip is not. Keep palette, lens feel and lighting consistent across all five.

3. 3

   Animate with small, specific motion

   Image-to-video models do best with one clear movement: a slow push-in, a hand placing the product, steam rising. Ask for 3 to 5 seconds. Avoid complex multi-person action and anything that depends on exact text.

   ✦Tip. Request a slightly longer clip than you need so you can trim the unstable first and last frames.

4. 4

   Review every clip frame by frame

   Scrub for warping, flickering hands, changing logos and objects that appear or vanish. Regenerate the clip or hide the problem with a cut. Never leave a product label that morphs.

5. 5

   Assemble in a normal editor

   Cut the clips to length in an editor, add captions you write yourself, and use music you have a licence for. Keep the first second clear and visually strong.

6. 6

   Export for the platform

   For vertical placements export 1080 × 1920 H.264 at the platform’s recommended frame rate and keep captions inside the safe zone. Watch the exported file on a phone before you post it.

7. 7

   Disclose where it is required

   Some platforms and some countries ask creators to label AI-generated or synthetic media. Read the current rules for each place you post and label accordingly.

Replace anything in [square brackets]. Prompts are starting points: results vary between tools and versions, so review every output.

Prompts are written in English because image, video and most text models follow English instructions best. To get answers in your language, add “Reply in English.” to a prompt for a text assistant.

Message: [one sentence]. Product: [what it is]. Write a five-shot, 15-second vertical video plan: hook, desire/problem, product in use, detail/benefit, call to action. For each shot give one drawable sentence, the camera move, and a still-image prompt. Avoid text inside shots and avoid complex crowd scenes.

Use the supplied image as the first frame. Slow push-in, subject stays centered and sharp, soft natural light stays constant, subtle background movement only. Keep the product, its label and its shape unchanged. 4 seconds.

Here are my captions for a 15-second video: [paste]. Rewrite each to five words or fewer, keep the meaning, and mark any claim that needs proof. Do not add statistics or promises.

• ✕Prompting one 15-second clip instead of five short shots.
• ✕Trying to render legible text inside the video.
• ✕Skipping the frame-by-frame review.
• ✕Using unlicensed music because “everyone does”.