How to Use ChatGPT Merchant Center: Full Setup Guide

Editorial Team Sep 22, 2026
How to Use ChatGPT Merchant Center: Full Setup Guide

This review is researched from each provider's official pricing, plans and public user feedback — see our editorial process for how we keep it accurate.

What is ChatGPT Merchant Center?

ChatGPT Merchant Center is OpenAI's system for letting online retailers feed their product catalog directly into ChatGPT's shopping experience. Instead of crawling your storefront the way a search engine would, you push a structured product feed — prices, stock levels, images, shipping details — straight to OpenAI on a schedule you control. That feed is what lets ChatGPT recommend your products by name, with accurate pricing, when someone asks it for shopping advice.

For merchants, this is the closest thing to "SEO for ChatGPT" that currently exists: your products can only appear in ChatGPT's shopping answers if OpenAI has a feed to pull from, and getting that feed right is most of the work.

How it fits into ChatGPT Shopping

ChatGPT's shopping features cover two separate jobs, and it helps to keep them apart:

  • Product discovery — ChatGPT surfaces relevant products when someone asks a shopping question ("best running shoes under $100"). Powered entirely by merchant product feeds.
  • In-chat checkout — for merchants who integrate the Agentic Commerce Protocol (ACP), a buyer can complete part of a purchase without leaving the chat, with the merchant's systems handling payment and fulfillment behind the scenes.

Discovery is open to any merchant who submits a compliant feed. Checkout integration is a separate, more involved step layered on top — don't assume the two are the same program; you can get discovery working without touching checkout at all.

Who can apply

Merchant Center is aimed at businesses with a real product catalog — not individual sellers, dropship storefronts without inventory, or affiliate sites reselling other stores' products. You'll generally need:

  • A live e-commerce storefront with real, in-stock products
  • A registered business entity that can pass OpenAI's verification
  • The technical ability to generate and host a structured product feed (either in-house or through a partner platform)

Shopify merchants have the most direct path in, since Shopify has built native syndication tools that generate a compliant feed from your existing catalog without custom export code.

Step-by-step: setting up your product feed

Step 1 — Apply through the merchant portal

Start at OpenAI's merchant application page (linked from ChatGPT's shopping help documentation). You'll submit business details, storefront URL, and category of goods sold. OpenAI reviews applications before granting feed-ingestion access — this isn't a self-serve toggle.

Step 2 — Build your feed in a supported format

Once approved, you build your product feed as JSONL, CSV, TSV, or Parquet, matching OpenAI's published field schema. At minimum you'll need to map:

Field typeExamples
IdentifiersSKU, GTIN/UPC, product ID
Core contentTitle, description, category, brand
PricingCurrent price, currency, sale price if applicable
AvailabilityIn stock / out of stock / preorder, quantity
MediaPrimary image URL, additional image URLs
FulfillmentShipping cost, delivery estimate, return window

Accuracy matters more than completeness — a feed with fewer fields but correct pricing and stock status will outperform a feature-rich feed that's stale or wrong, since OpenAI's ranking weighs reliability heavily.

Step 3 — Submit a sample batch for validation

Before your full catalog goes live, you typically submit a sample of roughly 100 products so OpenAI's ingestion pipeline can validate formatting, required fields, and data quality. This catches schema errors early instead of letting a broken full-catalog feed fail silently.

Step 4 — Push your full catalog and automate refreshes

Once your sample passes, submit the complete catalog and set up an automated refresh cycle. OpenAI's spec supports updates as often as every 15 minutes, which matters most for fast-moving inventory. A feed that only updates once a day will show stale prices and "in stock" items that have already sold out — both hurt trust and, over time, ranking.

Product feed requirements at a glance

RequirementDetail
Accepted formatsJSONL, CSV, TSV, Parquet (compressed variants accepted)
Delivery methodSecure push to an OpenAI-provided endpoint — not a public URL OpenAI crawls
Refresh frequencyUp to every 15 minutes; daily minimum recommended for stable catalogs
Validation stepSample batch (~100 products) reviewed before full catalog goes live
Required dataIdentifiers, title, price, availability, at least one image

Exact field names and rules are published in OpenAI's developer docs and change as the program matures, so confirm current requirements against the official spec before building your pipeline.

Shopify merchants: the easier path

If your store runs on Shopify, you largely skip the manual feed-building work. Shopify has built direct syndication into its platform so a store's existing product catalog can be pushed to ChatGPT (and other AI shopping surfaces) without a custom integration. In practice this means:

  • Enabling the relevant sales channel or app from the Shopify admin
  • Shopify handling the field mapping and refresh cadence on your behalf
  • Your existing product data (titles, images, inventory) becoming the source of truth, rather than maintaining a second, separate feed

This is worth checking first if you're on Shopify — building a custom JSONL export pipeline is real engineering work you can likely skip.

Instant checkout and the Agentic Commerce Protocol

Beyond product discovery, OpenAI has also worked on letting shoppers complete purchases without leaving ChatGPT, built on the Agentic Commerce Protocol (ACP) — an open standard OpenAI developed with Stripe. ChatGPT collects the buyer's order details and renders a checkout interface, but payment processing, tax calculation, and fulfillment stay on the merchant's own systems; OpenAI has been explicit that it is not the merchant of record.

This checkout layer has moved quickly since it first appeared, and reporting on which merchants have it enabled — and under what name — has shifted more than once. Treat checkout integration as a separate, optional step from feed submission: get your product feed accepted and showing up in ChatGPT's answers first, then check current checkout eligibility on OpenAI's merchant portal before building on top of it.

How ranking works in ChatGPT shopping

Being in the catalog doesn't guarantee recommendations. Publicly available guidance points to a handful of consistent ranking factors:

  • Availability — in-stock items are favored over out-of-stock or ambiguous listings
  • Price competitiveness — accurate, current pricing that reflects what a buyer would pay
  • Data quality — complete titles, descriptions, and images that let the model match your product to intent
  • Merchant authority — being the maker or primary seller tends to outrank resellers and affiliates

None of this is a guaranteed formula OpenAI publishes in full — but the underlying principle is familiar from Google Shopping feed SEO: clean, accurate, frequently updated data wins over a bigger but stale catalog.

Common mistakes to avoid

  • Treating the sample batch as a formality — issues that barely pass validation get worse at full catalog scale; fix warnings, not just hard errors.
  • Setting refresh frequency once and forgetting it — a feed that isn't kept current on price and stock hurts buyer trust.
  • Assuming checkout access comes bundled with discovery access — they're separate capabilities with separate eligibility.
  • Skipping the official documentation — feed specs change; treat any third-party guide, including this one, as a starting point.

Is it worth setting up?

For a merchant with a real catalog and the ability to keep a feed accurate and current, yes — being discoverable inside ChatGPT's shopping answers is a distribution channel with no ad spend required, only the engineering time to integrate correctly. For a small store without the resources to keep a feed fresh, a stale or low-quality feed can do more harm than good, since inaccurate listings erode the trust that drives ranking. Shopify merchants have the lowest-effort path in and should check native syndication tools before considering a custom build.

Frequently asked questions

Is ChatGPT Merchant Center free to use?

There's no publicly advertised fee for submitting a product feed for discovery. Transaction-level fees, where they apply, are tied to checkout integrations, not basic feed submission — confirm current terms on the merchant portal.

Do I need a developer to set up a product feed?

If you're not on Shopify (or another platform with native syndication), yes — building a compliant JSONL/CSV feed and automating its refresh is a real technical task, not a form you fill out.

How long does merchant approval take?

OpenAI doesn't publish a fixed timeline, and approval times have varied as the program has scaled. Applying early with complete business verification details tends to reduce back-and-forth.

Can affiliate sites or dropshippers use Merchant Center?

The program is built around merchants selling their own real inventory, not affiliate listings for products sold elsewhere.

What happens if my feed has errors?

Feeds are validated before full catalog ingestion; hard errors block acceptance, while soft issues (missing optional fields, weak descriptions) may be accepted but hurt ranking.

Is this the same as Instant Checkout?

No — product feed submission covers discovery. Checkout integration via the Agentic Commerce Protocol is a separate, optional layer with its own eligibility requirements that have shifted since launch.

Will this replace my Google Shopping feed?

No — treat it as an additional channel alongside Google Shopping, marketplaces, and your own site. Accurate, current, well-structured product data benefits all of them.

#ai#ecommerce#guide#chatgpt

Read next