How to Turn Product Photos into Video Ads with AI

​

Step 1: Prepare your product photos

• Use clean, well-lit photos, ideally on a plain background.
• Include more than one angle if you can.
• Make sure labels and logos are sharp.
• If you already have a lifestyle photo you like, include it as a starting frame.

​

Step 2: Decide the basics

​

Step 3: Write a shot-by-shot brief

“15-second 9:16 ad for this cold brew can (photos attached). 0–2s: Hook. Extreme close-up of condensation sliding down the can, cold blue light. Sound: ice cracking. 2–6s: Camera pulls back to reveal the can on a sunny café counter; a hand picks it up. Voiceover: ‘Cold brew, smooth enough to drink black.’ 6–11s: Quick cuts: pouring over ice, a first sip, a smile. Upbeat acoustic music builds. 11–15s: Can centered on the counter, logo facing camera. Text on screen: ‘Find it near you.’ Voiceover: ‘Find it near you.’ Keep the can’s label and colors exactly as in the photos. Draft quality first.”

​

Step 4: Review the draft

• The first two seconds: would this stop someone scrolling?
• The product: does it look right in every shot? Check the label, color and shape.
• Motion: anything warped, glitchy or unnatural?
• Sound: is the voiceover clear and the music right for the brand?
• Text: spelled correctly and on screen long enough to read?

​

Step 5: Refine

“Keep shots 2–4. Replace the opening with a slow pour of cold brew over ice in slow motion. Make the music a little quieter under the voiceover.”

“Make three versions with different openings: condensation close-up, ice dropping into a glass, and the can being pulled from a fridge. Keep everything after 2 seconds identical.”

​

Step 6: Add captions and make every size

• Captions: ask Glif to add subtitles; most people watch social video with the sound off.
• Other platforms: “Make a 1:1 version for feed and a 16:9 version for YouTube.”
• Shorter cuts: “Make a 6-second version using the hook and the end card.”

​

Step 7: Finalize

​

Choosing a video model

​

Want a person in the ad?

​

More on glif.app

• Product Ad with Seedance 2.0
• Multi-Angle product shot
• Create UGC Style Content
• Product Videos
• TikTok Videos
• Caption Videos
• Kling 3.0 Prompting Guide

​

Agent Reference

​

Workflow card

​

Brief structure

1. Platform, aspect ratio, total length
2. Timestamped beats (0–2s, 2–6s, …) with subject, action, camera movement, sound and on-screen text
3. Voiceover lines in quotes
4. Music direction
5. End card text and CTA
6. Fixed product attributes (“label and colors exactly as in the photos”)
7. Draft or final quality

​

Request pattern

{
  "name": "compose_project",
  "arguments": {
    "prompt": "15-second 9:16 ad for this cold brew can. 0–2s: extreme close-up of condensation, cold blue light, sound of ice cracking. 2–6s: pull back to sunny café counter, hand picks up can; VO 'Cold brew, smooth enough to drink black.' 6–11s: quick cuts of pour, sip, smile; upbeat acoustic music. 11–15s: can centered, logo to camera, on-screen text and VO 'Find it near you.' Keep label and colors exactly as in the photos. Draft quality.",
    "asset_ids": ["img_PRODUCTXXXXXXXXXXXXXX"]
  }
}

• Confirm platform, length, number of versions and draft/final before calling; video is the most expensive output.
• Request hook variations in the same call; continue in the same project_id for edits, captions, resizes and finals.
• Full video generation requires the Pro plan or above (Starter has limited video).