AI Product Video: Amazon and Shopify, Start to Finish

Make an AI product video for Amazon or Shopify from one listing photo: the exact model to pick, what it costs, and the specs each platform needs.

Updated Oct 3, 2026 7 min read

Editorial illustration of a product box on a pedestal turning into a glowing film strip of video frames, clean studio lighting, minimal flat-design style, no text, no logos, no people

You can turn a single product photo into a 5-to-15-second demo clip without filming anything, and it works for both an Amazon listing video and a Shopify product page. The fast path is: shoot or pull one clean photo, animate it with an image-to-video AI model, then export it to match the platform's specs. Below is the exact workflow, which model to use for which job, what it costs, and the file requirements Amazon and Shopify actually enforce.

What "AI product video" means here

An AI product video is a short clip generated from a text prompt, a photo, or both — not footage edited together, but motion, camera movement and (often) sound generated frame by frame by a video model. For e-commerce, the useful version is image-to-video: you start from the listing photo you already have, and the model adds a slow rotation, a light sweep, a hand picking it up, or a zoom into a feature, while keeping the product itself unchanged.

This matters because most product-video failures with AI come from using a pure text-to-video model, which redraws the product from scratch and gets the label, shape or color wrong. The fix is starting from your real photo and using a model built to hold a reference steady, not reinvent it.

The specs you're actually building for

Amazon

SpecRequirement
File formatMP4 or MOV
Resolution1280×720 minimum, 1920×1080 recommended
Aspect ratio16:9
Frame rate23.976, 24, 25, 29.97 or 30 fps
File sizeUnder 500 MB for listing videos
AudioOptional on detail-page videos, required on video ad formats

According to Inreels' 2026 Amazon video spec guide, Amazon's spec calls for MP4 or MOV format, a 1920x1080 (1080p) minimum resolution, a 16:9 horizontal aspect ratio, and a frame rate of 23.976, 24, 25, 29.97, or 30 fps. For the templated video ad format specifically, SellerApp's advertising video guide notes the video spec requires a 16:9 or 16:10 aspect ratio, a maximum duration of 15 seconds, a file size under 500MB, required audio, and no letterboxing. There's no single hard length limit on a detail-page video, but a short, feature-forward clip performs better than a long one. For a first demo clip, aim for 1080p, 16:9, under a minute, with the product benefit visible in the first few seconds.

Shopify

Shopify's native product gallery is more forgiving on length but strict on file type. Per the Shopify Help Center, video length can be up to 10 minutes, video size up to 1 GB, video resolution up to 4K (4096 x 2160 px), and video file type must be .mp4, .mov, or .webm. In practice you don't need anywhere near that ceiling — a 6-to-15-second looping clip on the main image slot does the job, and shorter files load faster on mobile, which is where most Shopify traffic converts.

Keep the actual file small even when the platform allows more — a bloated video slows the page and that costs more sales than it gains.

How to make the clip: step by step

  1. Pick your source photo. Use the cleanest existing listing image — square or near-square, product centered, plain or simple background. A photo the model already understands well animates more reliably than a low-light phone shot.
  2. Decide the motion. A slow 360° turn, a push-in on a label or texture, a hand lifting the product, or a "light sweep" reveal are the most reliable prompts for e-commerce. Avoid asking for complex scene changes in the first try.
  3. Generate the clip from the photo, not from a blank text prompt, so the product stays accurate.
  4. Add sound if the platform wants it. Some models generate ambient sound or a soft voiceover in the same pass; otherwise leave it silent for Shopify.
  5. Trim and export to the aspect ratio and length your platform needs, then re-export at a lower bitrate if the file is larger than needed.
  6. Upload — to Seller Central for Amazon, or the product media gallery for Shopify.

Do it in Plurel

Open plurel.tech, sign in free with Google (no card), and go to the Image row under the composer for photo prep, then Video for the clip.

  • If the background is busy, open Remove BG (0.5 credits) and strip it to transparent or white.
  • If the photo is low-resolution, run it through Upscale (1 credit) to 4K before animating — a sharp source photo gives a sharper clip.
  • If you need the product in a different setting (a kitchen counter, a model's hand, a shelf), use Edit with an instruction like "place this product on a marble kitchen counter with morning light," priced the same as a picture (0.5–4 credits depending on the image model you pick).

Step 2 — Animate it

Open Animate, upload the (cleaned) photo, and choose a video model based on what you need:

  • MiniMax H3 (from 12 credits) — generates native 2K video with native stereo sound straight from a photo. According to NovoAds' breakdown of the model, MiniMax positions it for advertising, branding, e-commerce, product design, UI/UX and gaming, which makes it a strong default when you want sound baked in without a separate audio step.
  • Seedance 2.5 (from 30 credits) — the strongest option when you need the product itself to stay perfectly consistent across a longer clip (up to 30 seconds), useful for a multi-angle product story in one generation.
  • Seedance 2.0 Mini (from 5 credits) — cheapest way to test a motion idea before spending more on a final render; 4–15 seconds, photo or text in, sound included.
  • Veo 3.1 Fast (from 20 credits) — 8-second 1080p clips with realistic motion physics, a solid choice when the product needs to look handled or dropped realistically (a bottle pouring, fabric moving).
  • Kling 3.0 (from 55 credits) — the premium option for a cinematic hero shot, 15 seconds at 1080p with optional end-frame control if you want the clip to land on a specific final frame (e.g., the product plus logo).

Type a prompt describing only the motion and lighting — since the photo already defines the product, don't re-describe it:

Slow 360-degree rotation on a turntable, soft studio lighting,
subtle reflection on the surface below, no background change
Camera pushes in slowly toward the product label, shallow depth
of field, light catches the texture, 3 seconds total

You'll see the credit cost before you confirm generation.

Step 3 — Extend, edit or add sound after the fact

  • Clip too short? Extend continues it past the end at the same video-model price.
  • Need a small change — different background color, remove a reflection, add a hand? Video edit changes an existing clip by instruction instead of starting over.
  • Want to keep iterating on the same scene? Just reply in the chat — "make it darker," "same shot but from the left," "continue that clip" — it edits what was just made.

Step 4 — Export and check the spec

Download the clip from your chat history and confirm it matches the platform table above before uploading — Amazon wants 16:9 and 1080p, Shopify accepts up to 4K but a 6–15 second loop is plenty.

Troubleshooting

The product looks slightly different in the video than the photo. This is the most common issue with text-to-video or weak reference models. Switch to a model with stronger reference control, like Seedance 2.5 or MiniMax H3, and keep the motion prompt simple — big scene changes give the model more room to drift.

The clip has the wrong aspect ratio for Amazon. Generate in 16:9 from the start where the model supports it; cropping a square or vertical clip after the fact usually cuts off part of the product.

File is too large for a fast Shopify page load. Shopify's ceiling is generous, but that's a maximum, not a target — re-export at 1080p and a standard bitrate, which is almost always well under the limit for a 10-second clip.

Need audio for an Amazon video ad but the clip is silent. Pick a model that generates native sound, like MiniMax H3, rather than adding a separate audio track afterward — it keeps the sound synced to the motion.

One clean photo, the right image-to-video model, and a quick check against the platform's spec table is all it takes to get a product video live today.

Questions people ask

What length should an AI product video be for Amazon?
Amazon's listing video player has no fixed maximum, but most sellers keep the main demo clip short; recommended practice is to show the product feature within the first few seconds and keep total length under a minute, with 1080p resolution and 16:9 aspect ratio.
What video format does Shopify accept?
Shopify's product media gallery accepts MP4, MOV, or WebM files up to 10 minutes long, 1 GB in size, and up to 4K resolution.
Can AI generate a product video from just one photo?
Yes. An image-to-video model can animate a single listing photo into a short clip with camera movement, and some models like MiniMax H3 and Seedance 2.5 add native sound in the same pass.
Which AI model keeps a product looking accurate and not distorted?
Models built for reference control, such as Seedance 2.5 and MiniMax H3's omni-reference mode, are made to hold a product's shape, label and color steady across the clip rather than redesigning it.
Do Amazon product videos need sound?
For Amazon's video ad formats audio is required, but for a standard product detail page video it is optional; either way, a quiet ambient or voiceover track generated natively by the video model saves a separate editing step.
How much does it cost to make an AI product video in Plurel?
It depends on the model and length: Plurel Video and Veo 3.1 Fast start at 20 credits, MiniMax H3 starts at 12 credits, and Seedance 2.0 Mini starts at 5 credits, all billed from your credit balance (a plan includes credits every month).

Sources

Keep going