We are investigating a discrepancy between StoreKit 2 product metadata and the Apple purchase confirmation sheet in an iPad app distributed through internal TestFlight. The app offers one auto-renewable monthly subscription, configured for China mainland at CNY 6.80/month with a two-week introductory free trial. We would appreciate guidance on targeted diagnostics, without assuming that the cause is an Apple platform defect.
On an iPad (A16) running iPadOS 27.0.1, we independently verified the installed build and captured these public diagnostic fields during one product-loading invocation:
Storefront.current.countryCode = CHN
Product.priceFormatStyle.currencyCode = USD
Product.price = 0.99
Product.displayPrice = US$0.99
Product.subscription.introductoryOffer = nil
Product.subscription.isEligibleForIntroOffer = true
The product and subscription-group identifiers matched the configured identifiers. Product metadata and Storefront.current were read sequentially with asynchronous calls, so this is not an atomic StoreKit snapshot. The currency and offer fields above are logged values, rather than an inference from the custom UI.
For this build and product, the tester reported that the Apple system purchase sheet displayed a two-week free trial followed by CNY 6.80/month. We did not capture a screenshot of that system sheet, and no successful purchase or charge is being claimed.
The developer subsequently deleted the original product and created a replacement in the same subscription group. We saved and refreshed App Store Connect to check the replacement's China-mainland availability, CNY 6.80 monthly price, and two-week free-trial offer. Build 1.0 (5) uses the replacement identifier and is available in the existing internal TestFlight group.
The tester confirmed installing build 5 and supplied an App screenshot showing US$0.99/month without the trial overview. The screenshot itself does not contain the build number; that attribution relies on the installation report. We have not yet captured fresh raw Product fields for this replacement. The tester also reports that its system purchase sheet is normal, but has not supplied its exact current wording or a screenshot. We therefore keep this evidence separate from the build 4 log values.
Changing the ordinary Apple account under Media & Purchases did not change the reported App display. This was not a dedicated Sandbox Apple Account comparison. We have not established the replacement product's propagation state or the platform root cause.
The following condensed excerpts show the production path, not a standalone reproduction project. Validation, state handling, and transaction verification are omitted here:
// Product loading:
let values = try await Product.products(for: [id])
guard values.count == 1, let product = values.first,
product.id == id else {
throw SubscriptionClientError.invalidProduct
}
products[id] = product
// The UI model copies product.displayPrice directly.
// Later, in the purchase method:
guard let product = products[productID] else {
throw SubscriptionClientError.unavailable
}
return try resolvePurchase(try await product.purchase(confirmIn: scene))
Within a single completed load, the UI fields and cached purchase Product come from the same result and configured identifier. We do not convert the display price to USD or configure a separate CNY price for the system sheet. The trial overview requires an eligible two-week free-trial offer in the product data. Audited distribution artifacts did not contain a local StoreKit testing configuration or price fixture.
What diagnostics would help distinguish product-metadata/storefront behavior from application state handling? Is a dedicated Sandbox-account storefront comparison the appropriate next control? Are there documented refresh requirements or known TestFlight issues relevant to this discrepancy? The existing build already logs product identifiers, raw price, currency, offer structure, and eligibility, so we can collect a targeted replacement-product observation if useful.