Amazon shoppable video: 1 to 12 minutes, 5 GB, MOV or MP4

Amazon's shoppable video for a product detail page takes a .mov or .mp4 file, up to 1080p, at most 5 GB, and between 1 and 12 minutes long; you can attach up to 4 shoppable videos to one product. That minimum matters for AI video: no single Sume generation reaches a minute, so a compliant file is an assembled cut, not one clip. Everything below about Amazon comes from Amazon's product video page, read on 2026-10-03; Sume's side comes from the Video Router and Timeline 1.0 docs.

This is the listing video you upload in Seller Central, which is a different product from the ad formats already covered in our Sponsored Products and Sponsored Brands posts. Those have their own length rules; this one does not.

What does Amazon require for a shoppable video?

The page lists the file rules, the seller rules, and one display rule that decides whether the video is seen at all.

Why is one AI clip not enough?

Sume's video catalog is built for short shots. seedance-2.5 accepts 4 to 30 seconds at 480p, 720p or 1080p, wan-3.0 accepts 2 to 30 seconds, gemini-omni-flash-1.1 is 3 to 10 seconds, and every other catalog model is capped at 15 seconds, per the Video Generation docs. Amazon's one-minute floor therefore needs at least two of the longest clips, and in practice a planned sequence: a product overview, a how-to step, a close-up, a setup shot.

Plan the cut around what the type you pick promises. A How-to or Setup video that never shows the steps is a poor fit for its label, and the AI clip should show the product you ship, which is the same discipline as in our synthetic performer post.

How do I assemble a 64-second 1080p cut on Sume?

Generate eight 8-second shots with seedance-2.5 at 1080p, 16:9, then join them with Timeline 1.0. Every clip URL must already be a media.sume.com artifact; Timeline refuses off-host URLs. Timeline's default output is 1080x1920 portrait, so set output to 1920x1080 yourself for a landscape listing video. Output width and height must be even integers from 256 to 2160, which 1920x1080 satisfies.

The call below uses the unbilled plan endpoint, which compiles the timeline and returns duration_seconds, segment_count and billable_minutes without creating a job or reserving credits. When the plan looks right, post the same body to /v1/timeline-1.0/render with an Idempotency-Key header. Timeline bills $0.10 per output minute rounded up, so a 64-second render is two billable minutes.

import os
import requests

clips = [f"https://media.sume.com/artifacts/artf_demo{i}/clip.mp4" for i in range(8)]
body = {
    "audio": {"mode": "silence", "duration_seconds": 64},
    "video": [
        {"source_url": u, "start": i * 8, "duration": 8}
        for i, u in enumerate(clips)
    ],
    "output": {"width": 1920, "height": 1080},
}
r = requests.post(
    "https://api.sume.com/v1/timeline-1.0/plan",
    headers={"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"},
    json=body,
    timeout=60,
)
print(r.status_code, r.json())

How do I check the file before I upload it?

Run the finished MP4 through video inspect with frames: false. That returns probe facts only, so you can confirm the duration sits inside 60 to 720 seconds and read the resolution before Seller Central does. Then compare against the table: under 5 GB, MOV or MP4, 1080p or less. A silent spine is a placeholder; once the script is final, replace it with a voiceover so the video is not mute.

Two decisions are Amazon's, not Sume's. The video type is a label you choose at upload, and the fewer-than-six-images rule means a listing with a full image set will not show the video in the main media block. Amazon's page does not say where the video goes once a product has six or more images, so check the live page before you count on the placement. Amazon's own figure is that detail pages with their first shoppable video saw an average 24% increase in sales against pages without one, so treat the video as worth the assembly work, and measure your own listing rather than assuming it.