AI models · Ecommerce routing
All AI Models for Ecommerce Product Video
Listing modes stay on MiniMax H3; Prompt and Agent also pick Seedance 2.5, Wan 3.0, Veo 3.1, and Kling — route by channel, not hype. Each hub explains credits, QA, and when to upgrade.
MiniMax H3
Listing · WearableDefault for Product shot / Story / Wearable Ref2VA — gallery-safe masters.
Seedance 2.5
Prompt / AgentPick Seedance 2.5 in Prompt or Agent — cinematic beats with credits.
Wan 3.0
Prompt / AgentPick Wan 3.0 in Prompt or Agent — Alibaba motion with seller export.
Veo 3.1
Prompt / AgentPick Veo 3.1 in Prompt or Agent — cinematic Google Veo with credits.
Kling
Prompt / Agent · AdsKling in Prompt / Agent or Ad creative hooks from the SKU still.
Nano Banana
Still → listingNano Banana searches → SKU-true listing video (not official Nano Banana).
Seedance product video
Deep hubCredits, vs MiniMax H3, Shopify PDP — Seedance listing playbook.
Kling product video
Deep hubCredits, gallery vs ads, vs Runway — full Kling product-video playbook.
Runway
CompareCreative sandbox vs AmazVid listing workflow — when to use each.
Image to Video
All modesPhoto or URL → silent listing clip — route by channel.
NanoBanana product video
Deep hubAliases, i2v FAQ, still-prep vs listing motion.
OpenAI Video
CompareOpenAI video demos vs AmazVid listing workflow from stills.
Text to Video
IntentWhy ecommerce clips should start from a SKU still, not prompts alone.
FAQ — AI models on AmazVid
What AI models does AmazVid use for product video?
Listing modes and Story default MiniMax H3 (including Wearable Ref2VA). Prompt bar and Agent Video add a model picker: Seedance 2.5, Wan 3.0, Veo 3.1, Kling 3.0, and related models — credits scale by model. Compare hubs also cover Runway and NanoBanana-style still workflows.
Which model should I pick for Amazon gallery video?
Use MiniMax H3 Product shot for gallery-safe orbit motion. Reserve Wearable (H3) for fashion proof. Use Seedance / Wan / Veo / Kling in Prompt or Agent for cinematic or social beats — QA logos before any gallery upload.
Can I switch models without re-uploading the photo?
Yes. Paste a product URL once, then run Product shot, Wearable, Ad creative, or Prompt / Agent with a different model from the same still.
How is AmazVid different from Raphael AI or raw model playgrounds?
AmazVid is built for ecommerce listing workflows: URL ingest, first-frame lock, silent MP4 export, credits, and QA checklists — plus multi-model Prompt / Agent — not a vendor-only playground.