Hundreds of variations. One measured number each.
A POD catalog is a size-and-colour explosion — seven sizes times every design times every marketplace. AgentXray prices each child SKU from the settlement Amazon posted and lets your own AI draft the listing and ad changes, family by family.
- Per-SKU contribution from your settlement
- Listing drafts validated by Amazon before they go live
- We run a POD catalog ourselves
POD on reports: Forty-two children, one parent, and no idea which size pays.
The ads console attributes to the family. The fee lands per child. The printer's cost is in a third place. So a design that looks profitable at the parent can be losing money on three of its seven sizes.
"Is the 3XL worth keeping?" → nobody knows, so it stays.
POD with your own AI: Every child SKU priced from its own settlement line.
Referral, FBA and promotion amounts are read per child from the settlement events, matched to the cost you enter per size, and rolled up to the family. Your assistant can ask about one size, one design or one marketplace, and draft the listing or ad change for that family.
"Which sizes in the Brazil family lose money after fees?" → the list, with the number attached.
What a POD seller asks
Real tools on the connector, phrased the way a catalog seller thinks.
- Which child SKUs in this family have negative contribution after fees?
- What did each design actually earn across US and CA last month?
- Draft a title under 75 characters for this parent and validate it against Amazon.
- Which of my listings are buyable but not discoverable right now?
- Where is ad spend going to a size that never sells?
- How many days of cover does the best-selling size have?
- Did last month's price ladder change help net contribution?
Listing changes, validated before they go live
A POD catalog lives on its listings. Title, bullets, keywords and images are drafted by your assistant and sent to Amazon for validation before anything is applied — a title over 75 characters or a bad main image is caught in the draft, not after the listing is suppressed. You approve; then it goes live.
Who this is for
Apparel and jersey catalogs
Multi-size, multi-colour, multi-design families on US and CA. Per-child fees differ by size band; the roll-up hides it. This surface does not.
Designers scaling on Amazon
You add designs weekly and need to know, per design, which marketplace and which size actually pays before you add the next ten.
Sellers already on Claude or ChatGPT
You want to ask about a family in plain language and get a draft back, not open five reports.
What we do not do — said up front
- We do not generate print files, pick carriers or print shipping labels. AgentXray is the measurement and drafting layer, not the print shop.
- We do not access customer names or addresses. Amazon's PII role was not granted to us and we did not build around it; fulfilment stays with your print partner.
- We do not estimate a child's fee from its parent. Each child is priced from its own settlement line, or shown as a gap until that line posts.
- Our own POD catalog is small: twenty-seven live jersey listings on US and CA. It is how these rules were learned, not a claim about scale.
Questions
Does it work with Printful or another print partner?
The measurement layer reads what Amazon posted — fees, orders, settlements — regardless of who prints. You enter your cost per size once; fulfilment stays where it is.
Can it handle variation families?
Yes. Parent and child SKUs are read as a family; contribution is measured per child and rolled up. Listing drafts target the parent and validate the 75-character title rule and the child structure before submission.
Which marketplaces?
US, CA, UK, DE and AU today. Our own POD catalog runs on US and CA, which is where the family rules were learned end to end.
Can the AI change my listings on its own?
No. It drafts the change, Amazon validates it, and a person approves it in the browser. The connector cannot authenticate against the approval endpoint.