ChatGPT shopping cards schema - cloro
Shopping Cards Data Documentation
This section documents the shopping cards data returned by the ChatGPT endpoint, extracted when ChatGPT returns product or commercial information. They are opt-in: set include.shopping: true, which adds +2 credits to the base cost. For use cases, pricing context, and copy-paste examples, see the ChatGPT shopping API page on the product site.
{
"prompt": "What are the best sneakers under $100?",
"model": "CHATGPT",
"country": "US",
"include": {
"shopping": true
}
}
Model availability Shopping cards mostly appear when ChatGPT uses the gpt-5-3-mini model. ChatGPT selects the model per request, so cards may be absent even with the flag set.
Shopping Card Structure
Each shopping card contains:
| Field | Type | Description |
|---|---|---|
tags |
array | Category tags for the shopping card (e.g., ["electronics", "deals"]) |
products |
array | Array of product information objects |
Category and attribute data There is no dedicated category field or generic attributes bag. Category signals live in the card’s tags (high-level, e.g. ["electronics", "deals"]) and each product’s featured_tag (product-level). Attributes are individual typed fields — title, price, rating, merchants, offers[] and so on.
Product Information
Each product includes the following e-commerce data:
| Field | Type | Description |
|---|---|---|
title |
string | Product name |
position |
integer | 1-indexed rank across all products in all shopping cards in the response (flat, not reset per card). |
url |
string | Product page URL with ChatGPT attribution |
description |
string | Product description |
price |
string | Current price (e.g., "$57.00") |
featured_tag |
string | Product category or style tag |
merchants |
string | Merchant names (e.g., "Amazon.in + others") |
image_urls |
array | Array of product image URLs |
rating |
float | Product rating (0-5 scale) |
num_reviews |
integer | Number of reviews |
id |
string | Unique product identifier |
cite |
string | Citation reference identifier (e.g., "turn0product0") |
offers |
array | Array of shopping offers from different merchants |
offers_see_more_boundary |
integer | Number of offers shown before “see more” |
rating_grouped_citation |
object | Rating source information with supporting websites |
showcase_metadata |
object | Product showcase display metadata with image positioning |
providers |
array | List of data provider identifiers |
product_lookup_key |
object | Key used to look up product updates |
product_lookup_data |
object | Structured product lookup query data |
show_price_disclosure |
boolean | Whether to show price disclosure |
analytics_meta |
object | Analytics metadata with product_event_uuid |
metadata_sources |
array | Sources of product metadata (e.g., ["p2"]) |
generatedProductQuery |
string | Per-product shopping fan-out query... |
Offer Information
Each offer contains merchant-specific details.
Finding a product’s brand There is no top-level brand field. Brand is exposed as offers[].brand, but ChatGPT populates it on only a subset of offers, so it is frequently null or absent. When no offer carries it, parse product.title — the brand is usually the first token or two (e.g. "Adidas VL Court 3.0"). Fall back to merchants only when the merchant name matches the brand (e.g. "adidas + others").
| Field | Type | Description |
|---|---|---|
merchant_name |
string | Merchant name (e.g., "Amazon.in", "Macy’s") |
merchant_subtitle |
string | Additional merchant information |
marketplace_seller |
string | Third-party seller on the marketplace |
seller_name |
string | Seller name |
brand |
string | Product brand |
product_name |
string | Product name as listed by merchant |
url |
string | Offer URL with ChatGPT attribution |
price |
string | Offer price (e.g., "$57.00") |
details |
string | Stock and delivery information |
original_price |
string | Original price before discount |
available |
boolean | Offer availability status |
checkoutable |
boolean | Whether offer can be checked out directly |
checkout_payload |
string | Checkout payload data |
checkout_image_urls |
array | Images for checkout flow |
is_digital |
boolean | Whether product is digital |
price_details |
object | Detailed price breakdown |
tag |
object | Promotional tag |
Price Details
Offers include detailed price breakdown:
| Field | Type | Description |
|---|---|---|
display_price |
string | Formatted display price |
base |
string | Base product price |
shipping |
string | Shipping cost |
tax |
string | Tax amount |
total |
string | Total price including any additional costs |
Response Example
{
"success": true,
"result": {
"text": "If you're shopping for sneakers under $100, here are some options...",
"shoppingCards": [
{
"tags": [
"stylish casual leather sneaker",
"heritage retro leather sneaker",
"canvas everyday skate-inspired"
],
"products": [
{
"title": "Adidas VL Court 3.0",
"position": 1,
"url": "https://www.adidas.com/us/vl-court-3.0-shoes/ID8797.html?utm_source=chatgpt.com",
"price": "$57.00",
"featured_tag": "stylish casual leather sneaker",
"merchants": "adidas + others",
"cite": "turn0product0",
"image_urls": [
"https://images.openai.com/static-rsc-1/AY2CiYu1..."
],
"id": "3250714974047560249",
"rating": 4.7,
"num_reviews": 10394,
"show_price_disclosure": false,
"offers_see_more_boundary": 3,
"providers": ["product_info"],
"metadata_sources": ["p2"],
"offers": [
{
"merchant_name": "adidas",
"seller_name": "adidas",
"product_name": "Adidas Women's VL Court 3.0",
"url": "https://www.adidas.com/us/vl-court-3.0-shoes/ID8797.html?utm_source=chatgpt.com",
"price": "$57.00",
"details": "In stock online and nearby, Delivery between Sat - Mon $4.99",
"original_price": null,
"available": true,
"checkoutable": false,
"is_digital": null,
"provider": "p2",
"price_details": {
"base": "$57.00",
"total": "$57.00"
},
"tag": {
"text": "Best price"
}
}
]
}
]
}
]
}
}