Hi everyone, I’m seeing a consistent one-day discrepancy between the expiresDate returned by the App Store Server API and the “Expires on” date shown in the iOS Settings / App Store subscription list. I’d like to confirm whether this behavior is expected or if I’m misunderstanding the way Apple rounds dates. Reproduction steps Step Action Result 1 Purchase a 1-month auto-renewable subscription on 23 June 2025 14:00 JST (UTC+9) Transaction succeeds 2 Immediately fetch the transaction with GET /inApps/v1/subscriptions/{transactionId} Response contains "expiresDate": "2025-07-23T05:00:00Z" (= 23 July 2025 14:00 JST) 3 On the same device open Settings › Apple ID › Subscriptions (or App Store › Account › Subscriptions) UI shows Expires on: 22 July 2025 The same happens for every monthly renewal and on multiple devices. Region is Japan, device time zone Asia/Tokyo. What I understand so far (and my hypothesis) Apple’s docs say a monthly subscription renews “on the same calendar date” of the next month, so renewal in this example is 23 July. If the renewal is scheduled for 23 July at 14:00 JST, the subscription is fully usable until the end of 22 July in calendar terms, because the new billing period starts the moment the 23rd begins in Apple’s canonical time zone. Therefore, it might be intentional for the UI to display 22 July—i.e., “you can keep using it through the 22nd; on the 23rd it renews.” This hypothesis makes sense internally, yet it still looks confusing to end users who read “Expires on 22 July” and assume access ends at 00:00 on the 22nd, a whole day earlier than in reality. Questions Is showing the day before the renewal date the official/expected behavior? If so, could Apple clarify that the “Expires on” label represents the last full calendar day rather than the exact expiry timestamp? Which value should we surface in-app when telling users “Your subscription is valid until …”? The server’s expiresDate (precise to the second, converted to user time zone), or A UI-style date that’s one day earlier, matching Settings / App Store? Does Apple have a public document describing this rounding/visual convention? Have other developers encountered user confusion about the apparent 1-day “shortening” and, if so, how did you word your in-app messaging? Any insight from Apple engineers or fellow developers would be greatly appreciated. Thank you!

Hi, We have a app with some auto-renewing subscription in a group of subscriptions. When a user upgrade from a subscription to another, the "user receive a refund of the prorated amount of their original subscription" (https://developer.apple.com/app-store/subscriptions/). How is the prorated calculated ? Example : subscription to 14,99$ / month. If subscriber upgrade after 10 days, is the refund calculated 10/30 of 14,99$ (so ~5$) ?

After the release of StoreKit 2.0, the in-app purchase failure rate increased by 63.19%, with the majority of errors being StoreKitError.unknown. When encountering this error, many users repeatedly attempt to make a purchase, but the outcome remains unchanged, resulting in the same unknown error. In some cases, users who wait approximately 2 minutes before retrying the purchase may either succeed or encounter the following error: “StoreKit.StoreKitError.systemError(Error Domain=NSCocoaErrorDomain Code=4097 "connection to service named com.apple.storekitd”)”. This issue has directly impacted our app's purchasing flow. Because our app only displays the promotional purchase offer once, these issues have significantly reduced the number of users successfully completing the offer. As a result, the conversion rate for this promotion has dropped well below expectations, negatively impacting our business metrics.

I'm using the iOS simulator with a StoreKit configuration file. I can see that there have been transactions while the app has been closed, but my StoreKit 2 listener is never called with those updates to be able to finish them When I open my app from a cold start. I've added a listener on application(_:didFinishLaunching:launchOptions:) like this: func startObservingTransactions() { task = Task(priority: .background) { for await result in Transaction.updates { if case .verified(let transaction) = result { await transaction.finish() } } } } But the Transaction.updates loop never gets called (have added breakpoints to check). It's only ever called when a purchase is made, or subsequent transaction renewals when the app is open. Only then it will get the previously unfinished transactions. Steps to reproduce: Create an app with a StoreKit config file (with sped up transactions) to purchase an item Make a purchase then quit the app Wait for a bit for more transactions to be made while the app is closed. Open the app from a cold start and none of the transactions will be finished by the listener in your app. Cancel the subscription via the transaction manager. Close and open the app from a cold start. The first transaction will be finished by the listener but none of the others will be. In Apple's docs it says If your app has unfinished transactions, the listener receives them immediately after the app launches Why is this not the case?

Yesterday I noticed that if I purchase a free trial subscription in my app using a sandbox account, when the subscription expires I see a “Billing Problem” message every time I open the app. ”allow purchases & renewals“ setting is ON so this shouldn’t happen. has anyone else seen this or knows how to resolve? observed on iOS 18.3.2 & 16.7.10 thanks

I am testing sandbox subscribe but I always recieve DID_FAIL_TO_RENEW I try to create new sandbox account and clear purchase history but not work can anyone help?

Hi all, I am adding the following StoreKit 2 code to my app, and I don't see anything in Apple's documentation that explains the unverified case. When is that case exercised? Is it when someone has tampered with the app receipt? Or is it for more mundane things like poor network connectivity? // Apple's docstring on `shared` states: // If your app fails to get an AppTransaction by accessing the shared property, see refresh(). // Source: https://developer.apple.com/documentation/storekit/apptransaction/shared var appTransaction: VerificationResult<AppTransaction>? do { appTransaction = try await AppTransaction.shared } catch { appTransaction = try? await AppTransaction.refresh() } guard let appTransaction = appTransaction else { AppLogger.error("Couldn't get the app store transaction") return false } switch appTransaction { case .unverified(appTransaction, verificationError): // For what reasons should I expect this branch to be entered in production? return await inspectAppTransaction(appTransaction, verifiedByApple: false) case .verified(let appTransaction): return await inspectAppTransaction(appTransaction, verifiedByApple: true) } Thank you, Lou

New subscriptions have been failing to renew in the sandbox for 3 days. I am seeing multiple posts and comments from people that appear to be experiencing the same issue. But I haven't seen any feedback from Apple representatives. I really do not want to launch a new app without seeing functioning renewals in the sandbox. Is there somewhere else we are intended to seek assistance?

Over the past two days, we've observed an unusual spike in requests from some iOS users to our server endpoint responsible for verifying App Store purchase receipts. After sampling and analyzing the data, we found that the cause is related to the behavior of StoreKit2.Transaction.updates. Specifically, when listening for transaction updates, the system returns a large number of historical transactions — some dating back as far as one year. These callbacks are interpreted as "new" transactions, which in turn trigger repeated calls to Apple’s receipt verification servers, leading to an abnormal surge in traffic and putting pressure on our backend services. This behavior is ongoing and is something we've never encountered in our previous experience. It appears to be outside of expected behavior, and we suspect it may be due to some kind of abnormality or unintended usage scenario. We would appreciate guidance on the following: Is this a known behavior or issue with StoreKit2? Are there specific device states or conditions that could cause the system to emit historical transactions in bulk? Are there any recommended practices for mitigating or filtering such transaction floods? We have attached logs for reference. Any help identifying the root cause or suggestions for investigation would be greatly appreciated. 2025-07-24 12:39:58.594 +0400 listenForTransactions :{ "appTransactionId" : "704289572311513293", "environment" : "Production", "inAppOwnershipType" : "PURCHASED", "originalPurchaseDate" : 1713445834000, "originalTransactionId" : "430001791317037", "purchaseDate" : 1713445834000, "quantity" : 1, "signedDate" : 1753346396278, "storefrontId" : "143481", } 2025-07-24 12:39:58.594 +0400 listenForTransactions :{ "appTransactionId" : "704289572311513293", "deviceVerificationNonce" : "c4f79de2-a027-4b34-b777-6851f83f7e64", "environment" : "Production", "inAppOwnershipType" : "PURCHASED", "originalPurchaseDate" : 1713445849000, "originalTransactionId" : "430001791317270", "purchaseDate" : 1713445849000, "quantity" : 1, "signedDate" : 1753346396278, "storefrontId" : "143481", "transactionId" : "430001791317270", } 2025-07-24 12:39:58.594 +0400 listenForTransactions :{ "appTransactionId" : "704289572311513293", "deviceVerificationNonce" : "02f305d7-0b2d-4d55-b427-192e61b99024", "environment" : "Production", "inAppOwnershipType" : "PURCHASED", "originalPurchaseDate" : 1713511999000, "originalTransactionId" : "430001792218708", "purchaseDate" : 1713511999000, "quantity" : 1, "signedDate" : 1753346396278, "storefrontId" : "143481", "transactionId" : "430001792218708", } 2025-07-24 12:39:58.598 +0400 [INFO] [MKPaymentService:23]: [XLPay] listenForTransactions :{ "appTransactionId" : "704289572311513293", "deviceVerificationNonce" : "5ca85907-1ab6-4160-828e-8ab6d3574d6f", "environment" : "Production", "inAppOwnershipType" : "PURCHASED", "originalPurchaseDate" : 1713512034000, "originalTransactionId" : "430001792219189", "purchaseDate" : 1713512034000, "quantity" : 1, "signedDate" : 1753346396278, "storefrontId" : "143481", "transactionId" : "430001792219189", } 2025-07-24 12:39:58.599 +0400 listenForTransactions :{ "appTransactionId" : "704289572311513293", "deviceVerificationNonce" : "04869b50-b181-4b69-b4ff-025175e9cf14", "environment" : "Production", "inAppOwnershipType" : "PURCHASED", "originalPurchaseDate" : 1713512049000, "originalTransactionId" : "430001792219440", "purchaseDate" : 1713512049000, "quantity" : 1, "signedDate" : 1753346396278, "storefrontId" : "143481", "transactionId" : "430001792219440", }

Dear Apple Technical Support Team, We have encountered a potential issue related to transaction handling while using StoreKit v2, and would greatly appreciate your assistance in confirming the behavior or providing any relevant guidance. Issue Description: When calling Transaction.unfinished and listening to Transaction.updates on the client side, we noticed that some transactions—despite having already been processed and successfully completed with finish()—are being returned again upon the next app launch, which results in duplicate receipt uploads. Current Handling Flow: 1. Upon app launch: • Iterate over Transaction.unfinished to retrieve unfinished transactions; • Simultaneously listen for transaction changes via Transaction.updates (e.g., renewals, refunds); 2. For each verified transaction, we immediately call await transaction.finish(); 3. We then construct a transaction model, store it locally, and report it to our backend for receipt verification; 4. After the server successfully verifies the receipt, the client deletes the corresponding local record; 5. On every app launch, the client checks for any locally stored receipts that haven’t been uploaded, and re-uploads them if necessary. Key Code Snippets: private static func verifyReceipt(receiptResult: VerificationResult) -> Transaction? { switch receiptResult { case .unverified(_, _): return nil case .verified(let signedType): return signedType } } public static func handleUnfinishedTransactions(payConfig: YCStoreKitPayConfig, complete: ((YCStoreKitReceiptModel?) -> Void)?) { Task.detached { for await unfinishedResult in Transaction.unfinished { let transaction = YCStoreKitV2Manager.verifyReceipt(receiptResult: unfinishedResult) if let transaction { await transaction.finish() if transaction.revocationDate == nil { let receipt = YCStoreKitV2Manager.createStoreKitReceiptModel( transation: transaction, jwsString: unfinishedResult.jwsRepresentation, payConfig: payConfig, isRenew: false ) complete?(receipt) } } } } } private func observeTransactionUpdates() -> Task<Void, Never> { return Task { for await updateResult in Transaction.updates { let transaction = YCStoreKitV2Manager.verifyReceipt(receiptResult: updateResult) if let transaction { await transaction.finish() if transaction.revocationDate == nil { let receipt = YCStoreKitV2Manager.createStoreKitReceiptModel( transation: transaction, jwsString: updateResult.jwsRepresentation, payConfig: self.payConfig, isRenew: false ) self.callProgressChanged(.receiptPrepared, receiptModel: receipt, errorType: .none, error: nil) } } } } } Our Questions: 1. Is it possible for Transaction.unfinished or Transaction.updates to return transactions that have already been finished? Specifically, if a transaction was successfully finished in a previous app launch, could it still be returned again during the next launch? 2. Are there any flaws in our current handling process? Our current sequence is: finish() → construct model → local save → report to server → delete after verification. Could this order lead to timing issues where StoreKit considers a transaction unfinished? 3. If we need your assistance in investigating specific user transaction records or logs, what key information should we provide? We greatly appreciate your support and look forward to your response to help us further optimize our transaction processing logic.

Hi there! Whenever I try to add a new Offer Code for my app's subscription, I get an unknown error. When I get to the last step of the "Create Offer for Codes" flow, I get an error that error reads "An error has occurred. Try again later." I have been getting this same error for over a week now, so any help figuring out how to add new offer codes would be greatly appreciated!

We’re testing SKAN postbacks via AdAttributionKit but aren’t receiving any requests on our server even after generating development impressions and triggering a postback. Setup: Domain: https://linkrunner-skan.com Configured in Info.plist as: <key>NSAdvertisingAttributionReportEndpoint</key> <string>https://linkrunner-skan.com</string> <key>AttributionCopyEndpoint</key> <string>https://linkrunner-skan.com</string> Apple automatically appends the .well-known paths: /.well-known/private-click-measurement/report-attribution/ /.well-known/skadnetwork/report-attribution/ ATS diagnostics for the domain: PASS for all tests (TLS 1.0–1.3, PFS disabled, arbitrary loads allowed, etc.) Both .well-known paths are publicly accessible and return 200 OK Testing Flow: Enabled Developer → AdAttributionKit Developer Mode on iOS (15+) Followed Apple’s official guide: Testing AdAttributionKit with Developer Mode Generated test impression using: createAdAttributionKitDevelopmentImpression implemented in SKANManager.swift Called Postback.updateConversionValue with lockPostback: true Created Development Postback from Developer Settings Waited 30+ minutes while intercepting server requests (proxy + backend logs) What We’ve Tried So Far: Confirmed ATS compliance with nscurl --ats-diagnostics (all PASS) Verified .well-known paths are accessible publicly without redirects Tested endpoints manually with a POST request – server responds 200 OK Confirmed Info.plist entries exactly match Apple’s required keys Double-checked iOS device is running iOS 15+ with Developer Mode enabled Repeated test flow multiple times with fresh impressions and postbacks Waited up to 1 hour for postback (in case of delays) Issue: No POST requests are being received from Apple to either .well-known endpoint, even though the setup appears correct and ATS tests pass. References Used: Configuring an Advertised App Generating JWS Impressions Question: Has anyone faced a similar issue where AdAttributionKit Development Postbacks are not firing despite correct Info.plist setup, ATS compliance, and reachable .well-known endpoints? Any insight into possible missing configuration steps or testing nuances would be greatly appreciated.

Hello, I’m experiencing an issue with StoreKit 2 when passing a new appAccountToken for each purchase request. Case-ID: 15948169 (for DTS reference) Description of the Problem When initiating a purchase, I generate a new UUID to use as the appAccountToken: let serverTransactionId = UUID() let options: Set<Product.PurchaseOption> = [ .appAccountToken(serverTransactionId) ] let result = try await product.purchase(options: options) Expected Behavior: Each new purchase should return the updated appAccountToken that I pass into the purchase options. Actual Behavior: The payload response after success always contains the same appAccountToken from the very first transaction. It ignores subsequent UUIDs I pass and keeps reusing the original one. This causes issues because the same identifier is being reused across multiple transactions, making it difficult to map purchases to the correct user session. Steps to Reproduce Generate a fresh UUID using UUID(). Pass it as .appAccountToken when calling purchase(). Complete the transaction in the sandbox environment. Inspect the payload response → The appAccountToken value is always the same as the first one used, not the newly provided one. Additional Info I do have a focused test project that reproduces this issue. The issue appears specific to appAccountToken persistence across multiple transactions. Has anyone else experienced this behavior with StoreKit 2? Is this expected (Apple caching the first token) or could this be a bug?

Hi, I'm reaching out to report a recurring issue with in-app purchases on iOS that seems to be related to Apple’s transaction handling — not to third-party libraries. In my Flutter application, I use both StoreKit2 and StoreKit1 (for comparison) via different packages, including the official in_app_purchase package. However, in both cases, I’m experiencing unexpected reuse of transactionId and appTransactionId values, even when initiating fresh purchases with unique appAccountToken values. Problem Summary: Purchase Stream Returns Old Purchases When calling buyNonConsumable() with a new product, the purchase stream still returns data for a previously purchased product, despite clearing all Sandbox transactions and using a new applicationUserName for each attempt. Transaction IDs Reused Across Distinct Purchases Even when generating a new UUID for appAccountToken on each purchase, the returned appTransactionId and transactionId are reused — this breaks our server-side logic, which expects these fields to uniquely identify purchases and users. Example Logs: // First purchase { "appAccountToken": "2d5a0880-f68e-44a7-a414-f51204e63904", "appTransactionId": "704464472748013865", "transactionId": "2000000928154716" } // Second purchase (different user context) { "appAccountToken": "2d5a0880-f68e-44a7-a414-f51204e63904", "appTransactionId": "704464472748013865", "transactionId": "2000000928429780" } Even when using a different productId, the appTransactionId stays the same. When using StoreKit1, the productId updates properly, but the transactionId still matches the previous one. This behavior also affects App Store Server Notifications (V2): we have observed notifications tied to appAccountTokens from completely different user accounts (based on internal logs), sometimes delayed by days or weeks. I’ve prepared a reproducible example using the official Flutter in_app_purchase sample with minimal changes — you can find it here: Github gist The code is almost identical to the package example. I only added UUID generation for applicationUserName in _getToken(). In the actual app (not in this example), I retrieve the token from an API. Additional Observations from the Community: We’ve also found similar issues reported in other frameworks and languages. For instance, a developer using react-native-iap observed that App Store Server Notifications in TestFlight were tied to previously deleted users, even after signing up with a new user account and generating a new appAccountToken. Details here: User A deleted → User B signs up → receives upgrade event with User A’s token Notification uses appAccountToken from old account, not the new one This strengthens the suspicion that the issue may be related to how Apple associates transactions with Apple IDs in test environments. Questions: Is it expected for transactionId or appTransactionId to persist across purchases within the same Apple ID, even for different user contexts (e.g., separate logins in the same app)? Is there any official recommendation for avoiding this kind of data reuse in Sandbox or TestFlight environments? Should I expect appAccountToken in server notifications to always match the latest value provided during the purchase? Thank you in advance for your assistance. I would appreciate any clarification or advice regarding this issue, as it impacts production logic that relies on these identifiers being unique and consistent.

Both the legacy StoreKit API and the new StoreKit 2 API return the incorrect storefront countryCode. My actual Apple ID region is Germany, and my Sandbox test user is set to France, yet the SDK consistently returns USA. Expected Results: The returned storefront countryCode should reflect the correct region - sandbox user region if signed in and real user region if not signed in with sandbox). Actual Results: Returned country code is USA with both SKPaymentQueue.default().storefront?.countryCode and await Storefront.current?.countryCode. Signing out/in, device reboot and even reset do not help, I'm stuck with USA storefront.

We offer a 3-day free trial, and our paywall clearly states that users will be charged after the trial ends. However, some users request refunds after the charge - even after fully using our app for days or even weeks. In some cases, refunds are approved despite the users having consumed our AI processing services for up to a month. Since our app relies on backend AI processing, each user session incurs a real cost. To prevent losses, we utilize RevenueCat’s CONSUMPTION_REQUEST system and have set our refundPreference to: "2. You prefer that Apple declines the refund". Until recently, Apple typically respected this preference, and 90% of refund requests were declined as intended. However, starting about a week ago, we observed a sudden reversal: Apple is now approving around 90% of refund requests, despite our refund preference. As a result, we are operating at a loss and have had to halt both our marketing campaigns and our 3-day free trial. We’re trying to understand whether this shift is due to a change in Apple’s refund policy, or if we need to handle CONSUMPTION_REQUEST differently on our end. Has anyone else experienced similar changes? Any insights would be greatly appreciated.

I am using the public url https://api.storekit-sandbox.itunes.apple.com/inApps/v1/notifications/jwsPublicKeys to get the jwks keys to verify the signed payload for store kit payments. I am checking Apple server notifications. const APPLE_JWKS_URL = "https://api.storekit-sandbox.itunes.apple.com/inApps/v1/notifications/jwsPublicKeys" // Apple JWK set (cached by jose) const appleJWKS = createRemoteJWKSet(new URL(APPLE_JWKS_URL)); const jwks = await appleJWKS(); logger.debug("Apple JWKS Keys: %O", jwks); // Log the keys if (!signedPayload) { // return res.status(400).json({ error: "Missing signedPayload" }); } // Step 1: Verify JWS (signature + payload) using Apple's JWKS const { payload, protectedHeader } = await jwtVerify( signedPayload, appleJWKS, { algorithms: ["ES256"], // Apple uses ES256 for signing } );

Hello! The localization isn't working when using SubscriptionStoreView. The app hasn't been published yet. The subscription has been created and localization strings have been added. Status - ready to submit. Testing environment: Sandbox When calling SubscriptionStoreView, the debug console shows this error: GenerativeModelsAvailability.Parameters: Initialized with invalid language code: ru-RU. Expected to receive two-letter ISO 639 code. e.g. 'zh' or 'en'. Falling back to: ru Despite this, the subscription interface appears in English when Russian is expected. I don't use any locale setting for ru-RU anywhere in my code. The test device's region is set to Russia, and the language is Russian. Any help would be appreciated.

Hello, Is anyone else seeing Purchase.PurchaseResult.UserCancelled, despite a successful transaction? I had a user notify me today that he: Attempted a purchase Entered payment credentials Was asked to opt in to email subscription notifications Opted In Was shown my app's "User Canceled Purchase" UI Attempted to repurchase Was alerted that he was "Already Subscribed" I have adjusted my code to check Transaction.currentEntitlements on receiving a .userCancelled result, to avoid this in the future. Is this logically sound? Here is my code - please let me know if you see any issues: func purchase(product: Product, userId: String) async throws -> StoreKit.Transaction { let purchaseUUID = UUID() let options: Set<Product.PurchaseOption> = [.appAccountToken(purchaseUUID)] let result = try await product.purchase(options: options) switch result { case .success(let verification): guard case .verified(let tx) = verification else { throw PurchaseError.verificationFailed // Show Error UI } return try await processVerified(tx) case .userCancelled: for await result in Transaction.currentEntitlements { if case .verified(let tx) = result, tx.productID == product.id, tx.revocationDate == nil { return try await processVerified(tx) } } throw PurchaseError.cancelled // Show User Cancelled UI case .pending: throw PurchaseError.pending // Show Pending UI @unknown default: throw PurchaseError.unknown // Show Error UI } } @MainActor func processVerified(_ transaction: StoreKit.Transaction) async throws -> StoreKit.Transaction { let id = String(transaction.id) if await transactionCache.contains(id) { await transaction.finish() return transaction // Show Success UI } let (ok, error) = await notifyServer(transaction) guard ok else { throw error ?? PurchaseError.serverFailure(nil) // Show Error UI } await transaction.finish() await transactionCache.insert(id) return transaction // Show Success UI } The only place the "User Cancelled Purchase" UI is displayed on my app is after the one instance of "throw PurchaseError.cancelled" above. This happened in Production, but I have also seen userCancelled happen unexpectedly in Sandbox. Thank you for your time and help.

We're experiencing an unusual issue with SKProductsRequest in our app's in-app purchases. Despite confirming that all our product IDs are correct, the invalidProductIdentifiers property of SKProductsResponse sometimes contains all the requested product IDs. Here are the specifics: The issue occurs randomly and is not persistent. For instance, around 2024-12-26 -8.0 01:06:04, this problem occurred 3 times in quick succession. The didFailWithError method of SKRequestDelegate is not triggered, reporting no errors. Some users report that after encountering this issue, it resolves itself after about an hour. However, other users have reported this problem persisting for several days. We have repeatedly verified the correctness of our product IDs. Given the intermittent nature of the problem, we suspect this might be an issue on the App Store server side. We're looking for suggestions to resolve or mitigate this issue to improve the reliability of in-app purchases in our application. Has anyone encountered a similar issue? Or does anyone have suggestions that could help us further diagnose and resolve this problem?