What App Review Looks For, as Far as We Know: October 2026

This post has not been translated yet, so you are reading the English original. Read it at the English URL

What does Apple’s App Review check, and what is it like to be on the other side of it? As far as the public record shows on October 5, 2026: a team Apple’s own site now sizes at “nearly 500 dedicated experts,” with machine learning alongside it, reviewed 9.1 million submissions in 2025 and rejected 2.09 million of them, nearly one in four; Apple says “On average, 90% of submissions are reviewed in less than 24 hours,” and that “over 40% of unresolved issues are related to guideline 2.1: App Completeness,” the rule about crashes, placeholder content, missing demo access and purchases that aren’t there.12345 Nobody outside Apple has published the reviewer’s checklist. What follows is the next best thing: Apple’s own pages and reports, the rejection messages developers have posted, the rule changes of the last fifteen months, and then our own app, Kiradex, walked through all of it against that record, with what we had to fix.

TL;DR

  • The numbers Apple publishes. 9.1 million submissions reviewed in 2025, over 2 million rejected (over 1.2 million of them new apps); over 443,000 for privacy, over 371,000 for copying, spam or misleading users, over 22,000 for hidden features. The transparency report puts the Performance section on 1.35 million of the 2.09 million rejections, Legal on 496,000, Design on 416,000, Business on 284,000.34
  • One week of percentages, eight years old, still the most recent guideline-level ranking Apple has given. For the week ending April 3, 2018: 2.1 App Completeness 42%, 4.3 Spam 10%, 5.1.1 Privacy 8%, 2.3.3 Accurate Metadata 8%, 5.1.5 Location 5%, 3.1.1 In-App Purchase 5%.6
  • Reviewers run iPhone apps on iPads, often enough to plan for it. Rejections have named the device: an iPad7,5 crash log in 2022; a developer’s account in May 2026 of fields clearing “on an iPad Air 11-inch (M3) running iPadOS 26.4.2.” Apple’s reply to one developer who argued the point: “all apps designed for use on iPhone must still be formatted correctly and behave properly when running on iPad.”7
  • The subscription rejection is a form letter, and it was still in use in September 2025. “We were unable to find the following required item(s) in your app’s metadata: – A functional link to the Terms of Use (EULA).” If you use Apple’s standard EULA, the fix is a link in the description.8
  • The rules moved in 2025 and 2026. Five age tiers (13+, 16+, 18+ added), social-media questions required from September 2026, random or anonymous chat pulled under guideline 1.2, a ban on other developers’ brands in your icon or name, and a submission model where the first in-app purchases and Game Center items of an app have to be added to the same submission as the version.9
  • Our own pass found eleven things to fix on a listing three people had read. Among them the Terms of Use link, prices in the product names, products and Game Center components that were not in the submission, a trademark in the keywords, and a server log that kept more than the privacy label said. The fixes are in the last section.

Where this comes from, and where it stops

Apple describes App Review in three places on its developer site: the App Review page, the App Review Guidelines (the page prints “Last Updated: June 8, 2026”), and App Store Connect Help; its consumer pages and its documentation add a sentence here and there, quoted below where they do.121011125 Once a year it adds numbers: a Newsroom release about fraud, each May so far, and the App Store Transparency Report, which counts submissions and rejections by guideline section.34 Rejection messages are not published by Apple, but developers post them, and the same sentences recur for years; the ones below are quoted from those posts with their dates, and they are the developers’ transcriptions, not Apple documents.8 What the team is like from the inside comes from court testimony and reporting, in the next section but one.

What this cannot tell you is what the reviewer assigned to your app will do on the day. Apple’s help says “Submissions may not be reviewed in the order you submit them,” and the guidelines say that “if your app is complex or presents new issues, it may require greater scrutiny and consideration.”102 Treat everything here as the shape of the process, not a promise about one review.

The scale of it

Apple’s May 2026 release says the App Review team “evaluated more than 9.1 million app submissions” in 2025 and “rejected over 2 million app submissions — including over 1.2 million new apps and nearly 800,000 app updates.”3 The transparency report for the same year gives the precise figures: 9,100,620 submissions, 2,093,244 rejected, 387,087 “approved after rejection.”4 That is 23.0 percent of submissions rejected, my arithmetic on Apple’s counts. It is a rate per submission, not per app: Apple’s footnote says “Apps may be submitted to App Review multiple times before they are approved,” and a rate per app cannot be worked out from what it publishes.4

The same report splits rejections by guideline section. In 2025: Performance 1,354,418; Legal 495,673; Design 415,532; Business 283,820; Safety 151,159.4 The sections add to more than the rejections because one submission can fail under several. Performance, the section that holds 2.1 App Completeness and 2.3 Accurate Metadata, appears on 64.7 percent of rejected submissions; it was 64.2 and 64.0 percent in 2023 and 2024. Legal is on 23.7 percent, Design on 19.9, Business on 13.6, Safety on 7.2.4 Apple’s own gloss on the page where it lists common issues: “On average, over 40% of unresolved issues are related to guideline 2.1: App Completeness, which covers crashes, placeholder content, incomplete information, and more.”1

The Newsroom releases add the themes Apple wants to talk about. For 2025: “over 22,000 submissions for containing hidden or undocumented features; over 371,000 submissions that copied other apps, were found to be spam, or otherwise misled users; and over 443,000 submissions for privacy violations,” and “nearly 59,000 apps” removed “for engaging in such bait-and-switch maneuvers.”3 The privacy figure has climbed each year: over 375,000 (2023), 400,000 (2024), over 443,000 (2025).3

On speed, Apple’s public pages say much the same thing in two wordings and date it nowhere: “On average, 90% of submissions are reviewed in less than 24 hours” on the App Review page, “We review over 90% of apps within 24 hours” on its developer-facing App Store page.15 On who does the work, Apple’s App Store page says today: “Every week, nearly 500 dedicated experts around the world review over 130K apps.”5 Its 2024 release had it as “Apple’s App Review team of over 500 experts,” reviewing “approximately 132,500 apps a week”; the 2025 release says “nearly 150,000 app submissions each week” with no headcount; and the 2025 total works out to about 175,000 a week.34 (The site’s own line read “over 500 dedicated experts … over 100K apps” in a June 2024 capture and “nearly 500 … over 130K” from an October 2025 capture on; I would not build a story on the difference between “over” and “nearly.”5) The 2026 release describes the division of labor: “Apple’s App Review has always combined human expertise with machine learning,” and “By utilizing AI to rapidly identify complex malicious patterns, analyze app similarity, and flag potentially problematic changes in app updates, Apple’s systems help human reviewers focus their expertise where it matters most.”3

Who looks, and for how long

The inside of the room is on the record because of a lawsuit. In Epic Games v. Apple, Apple’s head of App Review testified, its former head was deposed, Epic put documents from Apple’s own fraud team before the court, and the judge wrote findings of fact. Add two pieces of reporting from 2019, and what the former head has said since in an interview and in a post of his own (he says he built the team; he joined in 2009), and this is what the public record says about the people who review your app.131415161718

How many. The court’s finding, September 2021: “Part of this issue stems from the sheer number of apps submitted with only 500 human reviewers.”13 The Verge, reporting the trial’s fifth day that May, had heard it the same way: “only 500 people actually do the ‘human review.’”18 Trystan Kosmynka, Apple’s senior director of App Review, put the volume in a sworn declaration a month later: “Apple receives approximately one hundred thousand submissions of apps and app reviews per week, with a yearly total of over 4 million submissions. In 2019, for example, Apple received 4,808,685 submissions. Of those submissions, 1,747,278 were rejected … This amounted to a rejection rate of 36 percent in 2019 … In 2020, the rejection rate rose to around 40 percent.”14 Apple’s own May releases pick the story up from there: the 2024 release, on 2023, “over 500 experts” and 132,500 a week; the 2025 release, on 2024, nearly 150,000 a week; and 9.1 million submissions in 2025, of which the transparency report counts 23 percent rejected (a different document from the declaration, and not shown to count the same way); its site today says “nearly 500.”345 Before the lawsuit, Apple’s 2019 “Principles and Practices” page described “a combination of automated systems and hundreds of human experts. This team represents 81 languages across three time zones,” with “100 K reviewed” a week, “60 % approved … 40 % rejected,” and “The most common reason for rejection is minor bugs, followed by privacy concerns.”5 CNBC’s 2019 reporting, from former reviewers and “people familiar with the process,” had the team at “more than 300 reviewers” in “a pair of offices in Sunnyvale, California,” Apple employees “paid hourly” with badges and benefits, not contractors.15 Apple’s job postings, read this autumn, still put the App Review Specialist roles in Sunnyvale, by language (Spanish, French, Chinese, Turkish); the Spanish posting lists two minimum qualifications, fluency in English and Spanish and a forty-hour week, and all four set that week, Tuesday to Saturday for some roles and Sunday to Thursday for others, which is how a queue is staffed seven days.19

For how long. The one per-app average on the record comes through Epic’s pretrial filing, citing Phillip Shoemaker’s deposition: as of April 2016, it says, “the human review process typically took approximately 13 minutes per app for new apps and 6 minutes per app for updates to existing apps.”16 The same filing, citing a trial exhibit (PX06): “App Reviewers typically review between 50 to 100 apps per day, and productivity is tracked internally,” and, in the paragraph after it, that reviewers in certain instances “took less than a minute.”16 CNBC’s sources said the same about the quota: “Reviewers have daily quotas of between 50 and 100 apps,” the number each gets through in an hour tracked “by software called Watchtower,” per screenshots the network saw, and “Most reviewers spend only a few minutes per app, but many apps are simple …”15 Shoemaker, who ran the team from 2009 to 2016, told Bloomberg in 2019 that in the early days three reviewers looked at each app; in Bloomberg’s account of his remarks, review times came down when the process went to one set of eyes, and reviewers would come in each morning and pick 30 to 100 apps from a web tool.17 The arithmetic on Apple’s current numbers, mine not Apple’s: 130,000 a week across nearly 500 people is about 260 each, or 52 a working day, at the bottom of the quota CNBC reported. Apple’s July 2026 posting for a “Tools and Automations Lead, App Review” says where the tooling is going: “designing and deploying GenAI and agent-based solutions (LLMs, RAG, agents) that augment review, surface risk, and reduce manual effort.”19

What a human does, by the court’s account. The findings describe the human part plainly: “During app review, a human reviewer confirms that an app corresponds to its marketing description. … The human reviewer also checks that the app’s entitlements are reasonable for the task it purports to accomplish. Thus, a Tic-Tac-Toe game may be rejected if it asks for camera access or health data. Last, although not directly related, app review checks for offline safety issues.”13 Kosmynka’s declaration lists the methods without an order: “Apple performs computerized static and dynamic analysis, as well as manual human review, on the compiled software, or ‘binary’ file,” together with the metadata; 9to5Mac’s summary of his trial filing put the automated analysis first and the human review after it.1418 Epic’s filing, citing a trial exhibit (PX140), says the “guiding principle in App Review” has from its inception been that it “review[s] … not test[s]” apps.16 And on the equipment, CNBC: “Reviewers ‘claim’ a batch of apps through a web portal on a Mac desktop, called App Claim. They often examine the app on an attached iPad, even if it’s an iPhone app.”15 That sentence, from 2019, is the other half of the iPad rejections developers keep posting.

How fast, and what happens when it isn’t. The court, from the testimony: “Apple promises in its Service Level Agreement to complete a review of an app quickly: 50 percent within 24 hours and 90 percent within 48 hours. Apple claims that it is completing 96 percent of the reviews within 24 hours.”13 CNBC described the same metric from the inside with a looser target, “Apple aims for 50% of apps to be reviewed within 24 to 48 hours,” and one bad day for it: “On July 30 last year [2018], the SLA fell to 6%, according to a staff email. ‘Until we catch up, we are opening up 12-hour days.’”15

Mistakes. Kosmynka, on the stand: “It’s a human process. We do make mistakes … But we certainly try to rectify those mistakes when we learn of them.”18 The court: “the overall error rate appears to be relatively small, with Apple’s former head of app review testifying that it was around 15% in 2015,” without saying what the 15 percent measures, and with a warning in the footnote that rejection-rate statistics “say[] nothing about the error rate.”13 Apple’s own fraud-engineering team was harsher, in Epic’s pretrial account of its head’s deposition and of documents Epic cited: the head’s view was that App Review “would not accomplish anything that would deter a sophisticated attacker,” and the team, Epic says, “further believed” that “App Review is bringing a plastic butter knife to a gun fight” and that the process amounts to “a wetware [i.e., a human-led] rate limiting service and nothing more.”16 Shoemaker, in a September 2026 post ahead of a book: “An approval is not a certification that an app is safe. It never was. It is a record that the app behaved itself during the window when a human being was watching.”17

Escalation. Two bodies sit above the reviewer. The App Review Board hears appeals; CNBC’s sources said it “is partially composed of reviewers with good track records.” Above it the Executive Review Board, led in 2019 by Phil Schiller, took “difficult approval decisions” and the “publicly sensitive” ones; Kosmynka’s declaration describes it as the body “which creates policies that apply to Apple’s App Store and app review process,” and Shoemaker described its weekly meeting as “four hours in a room arguing about apps.”151417 Apple’s 2019 page said it in its own words: “All developers have the opportunity to have a rejection reviewed by the App Review Board. And the App Review team makes about 1,000 calls a week to developers to help them diagnose and resolve any issues that led to rejection.”5 Per 9to5Mac’s summary of Kosmynka’s trial filing, “less than 1% of developers appeal a rejection.”18

None of this is current beyond what Apple itself has said since: the headcount is from 2021 and from Apple’s site, the minutes from 2016, the quotas and the iPad from 2019, the two boards from 2019 to 2023, the shifts and the tooling from this year’s postings. It is the only inside view on the public record that I found.

What the review covers

Nobody at Apple has published the sequence a reviewer follows, and I am not going to invent one. What the record does give is the set of things reviews have been seen to check, because the rejection messages name them. Grouped, not ordered:

The listing. Metadata rejections cite the exact field: a support URL that “does not properly navigate to the intended destination,” screenshots that “do not sufficiently reflect your app in use,” an app name that “includes keywords or descriptors, which are not appropriate for use in an app name.”8 Apple’s statuses separate this outcome from a rejection of the binary: “Metadata Rejected” means “App Review didn’t accept your metadata. Read the message from App Review, edit the metadata to resolve the issue, and reply to the message from App Review,” and its help adds that “If your app was rejected for a metadata issue, you can resubmit the same build after resolving the issue.”11

The app on real devices, including an iPad. Guideline 2.4.1 says “iPhone apps should run on iPad whenever possible,” and the record shows reviewers holding apps to it. A 2022 rejection attached a crash log with "modelCode" : "iPad7,5" for an app its developer had tested in the simulator. A 2017 message: “We noticed that your app did not run at iPhone resolution when reviewed on iPad running iOS 10.2.” In May 2026 a developer reported four rejections because, in their words, typing into the sign-in fields “on an iPad Air 11-inch (M3) running iPadOS 26.4.2 causes the fields to clear.” When one developer sent an iPhone screen recording to argue the point, the reply was: “While we appreciate that your app is intended primarily for use on iPhone, in order to bring your app into compliance with App Store guidelines, all apps designed for use on iPhone must still be formatted correctly and behave properly when running on iPad.” The same thread records: “We are unable to provide further information regarding our review environment.”7 Four posts across nine years do not prove every review uses an iPad. They do show it is a normal thing for a review to do.

Launch, and whatever is reachable. Guideline 2.1(a) is the rule being applied: “Submissions to App Review … should be final versions with all necessary metadata and fully functional URLs included; placeholder text, empty websites, and other temporary content should be scrubbed before submission. … We will reject incomplete app bundles and binaries that crash or exhibit obvious technical problems.”2 The crash message has read the same way since at least 2015: “We were unable to review your app as it crashed on launch. We have attached detailed crash logs to help troubleshoot this issue.”8 Anything behind a login needs the demo account from App Review Information; the message for a bad one is “We were unable to sign in to review your app with the demo account information you provided.”8

Declared capabilities. Background modes get exercised: “Your app declares support for audio in the UIBackgroundModes key in your Info.plist, but we were unable to play any audible content when the app was running in the background.”8 Frameworks get checked against features: “your binary includes the PassKit framework for implementing Apple Pay, but we were unable to verify any integration of Apple Pay within your app.”8 Permission prompts are read for their purpose strings: “all permission request alerts need to specify how your app will use the requested information. … Make sure the purpose string includes an example of how the user’s data will be used.”8

Purchases, in the sandbox. Apple’s receipt-validation documentation says the review environment is the sandbox: use the test URL “when testing your app in the sandbox and while your app is in review,” and the production-first, fall-back-on-21007 rule is there because, in Apple’s words, “This approach ensures you don’t have to switch between URLs while your app is in testing, in review by App Review, or live in the App Store.”20 (The page is about receipt validation; it is still the place where Apple says which environment the review runs in.) Every product you configured is expected to be findable; 2.1(b): “If you offer in-app purchases in your app, make sure they are complete, up-to-date, visible to the reviewer and functional. If any configured in-app purchase items cannot be found or reviewed in your app, explain the reason in your review notes.”2 A 2024 rejection: “We are unable to complete the review of your app because one or more of your in-app purchase products have not been submitted for review.”8 For subscriptions the sign-up screen is held to the agreement, Schedule 2 §3.8(b): “Title of auto-renewing subscription,” “Length of subscription,” “Price of subscription, and price per unit if appropriate,” and “Links to Your Privacy Policy and Terms of Use must be accessible within Your Licensed Application.”21

Links. Support URL, privacy policy, Terms of Use, inside the app and in the listing. Apple’s common-issues page: “All links in your app must be functional. A link to user support with up-to-date contact information and a link to your privacy policy is required for all apps.”1 The 1.5 rejection names the URL that failed and asks for “a webpage with support information”; a page that was only an email address got the same message.8

The privacy policy, the privacy label, the age rating. These are rules more than observed checks: I found no posted rejection for a wrong privacy label or a wrong age rating. Guideline 5.1.1(i) requires the privacy policy “in the App Store Connect metadata field and within the app in an easily accessible manner.”2 For the label, Apple’s definition of collecting is the thing to know: “‘Collect’ refers to transmitting data off the device in a way that allows you and/or your third-party partners to access it for a period longer than what is necessary to service the transmitted request in real time,” and “if data is sent to your servers then immediately discarded after servicing the request, you do not need to disclose this in your answers in App Store Connect.”22 For the rating, 2.3.6: “Answer the age rating questions in App Store Connect honestly so that your app aligns properly with parental controls.”2

The decision, in a message. Apple’s help describes the exchange that used to be called the Resolution Center and is now the App Review section: “When your app is rejected, you’ll get a message containing information about the rejection, including how it violates the App Review Guidelines … You can correspond with Apple, and include attachments, such as screenshots and supporting documents, until you resubmit to App Review.”11 If a submission holds several items, “App Review needs to approve all items. To continue with the accepted items, you need to either edit and resubmit, or remove the rejected items from the submission.”11

Two process facts shape how you plan around all this. “Each platform can have one app version submission under review at a time.”10 And an app already on the store gets a concession: “bug fixes will not be delayed over guideline violations except for those related to legal or safety issues,” which you accept, in the App Review page’s words, by replying “to the offer message in App Store Connect.”21

What gets apps rejected

The most recent guideline-by-guideline ranking Apple has published is a single week, the seven days ending April 3, 2018, on a page it took down in 2019 (it had published weeks in 2015 and 2016 under the old numbering):6

Share Guideline
42% 2.1 Performance: App Completeness
10% 4.3 Design: Spam
8% 5.1.1 Privacy: Data Collection and Storage
8% 2.3.3 Performance: Accurate Metadata
5% 5.1.5 Privacy: Location Services
5% 3.1.1 Business: In-App Purchase
3% 2.3.7 Performance: Accurate Metadata
3% 4.2 Design: Minimum Functionality
2% 3.2.1 Business: Other Business Model Issues
2% 2.4.1 Performance: Hardware Compatibility

The top ten were 88 percent of rejections that week. The section-level counts in the transparency reports measure something different (citations per year, overlapping, not shares of one week), so they can’t confirm that list; what they say for 2023 through 2025 is that Performance sits on about two thirds of rejected submissions, Legal on about a quarter, Design on between a sixth and a fifth, Business on about one in nine to one in seven, Safety on about one in seventeen to one in fourteen.4 I found no vendor dataset from 2023 to 2026 that ranks rejections by guideline with a stated sample, though the search tools available to me were limited; the guides from RevenueCat, Runway and Bitrise are checklists, and one of them reproduces a template bug from Apple’s own page (a literal $!{infoPlistKey} where a key name should be).23

The exact words

Rejection messages are templates, and knowing the template tells you what was checked. These are quoted as developers posted them, with the year. Only the subscription and iPad messages are confirmed in use after 2024; Apple may have reworded the others since.8

  • Subscriptions (3.1.2), 2022, and quoted again in a September 2025 answer: “We noticed that your app did not meet all the terms and conditions for auto-renewing subscriptions, as specified in Schedule 2, section 3.8(b) of the Paid Applications agreement. … We were unable to find the following required item(s) in your app’s metadata: – A functional link to the Terms of Use (EULA) … If you are using the standard Apple Terms of Use (EULA), you will need to include a link to the Terms of Use in your App Description. If you are using a custom EULA, add it in App Store Connect.” The 2023 variant asks for the privacy link too, “in your app’s binary” and “in your app’s metadata.”
  • Crash (2.1), 2022: “We were unable to review your app as it crashed on launch. We have attached detailed crash logs to help troubleshoot this issue.”
  • In-app purchases not in the submission (2.1), 2024: “We are unable to complete the review of your app because one or more of your in-app purchase products have not been submitted for review.”
  • Purpose string (5.1.1), 2020: “We noticed that your app requests the user’s consent to access the location, but doesn’t sufficiently explain the use of the location in the purpose string.”
  • Support URL (1.5), 2019: “The support URL specified in your app’s metadata does not properly navigate to the intended destination.”
  • Screenshots (2.3.3), 2018: “We noticed that your screenshots do not sufficiently reflect your app in use. Specifically, your screenshots do not display the actual app in use.”
  • Minimum functionality (4.2), 2019: “We found that the usefulness of your app is limited by the minimal amount of content or features it includes. … We understand that there are no hard and fast rules to define useful or entertaining, but Apple and Apple customers expect apps to provide a really great user experience.”
  • Spam (4.3), 2017: “We noticed that your app provides the same feature set as other apps you’ve submitted to the App Store; it simply varies in content or language, which is considered a form of spam.” With a warning that “Future submissions of this app, and other apps associated with your Apple Developer account, will also experience a delayed review.”
  • User-generated content (1.2), 2017: “Your app enables the display of user-generated content but does not have the proper precautions in place,” followed by five requirements, two of which are stricter than the guideline’s text: “Require that users agree to terms (EULA) and these terms must make it clear that there is no tolerance for objectionable content or abusive users” and “The developer must act on objectionable content reports within 24 hours by removing the content and ejecting the user who provided the offending content.”
  • Background modes (2.5.4), 2023: “Your app declares support for audio in the UIBackgroundModes key in your Info.plist, but we were unable to play any audible content when the app was running in the background.”
  • Intellectual property (5.2), 2021: “Your app contains content or features that may violate the rights of one or more third parties. … please attach documentary evidence in the App Review Information section in App Store Connect evidencing that you have all necessary rights or permissions …”

What changed in 2025 and 2026

Dated from Apple’s developer news feed, newest first.9

  • August 5, 2026: a new “creative assets” type for product pages, where “you can go beyond in-use visuals to highlight your brand.” How App Review treats these against 2.3.3’s “app in use” rule for screenshots is not yet documented.
  • July 9, 2026: the age-rating questionnaire gains social-media questions. “A social media capability is defined as the ability to redistribute, amplify, or interact with user-generated content through a social feed or similar discovery method,” and “beginning in September 2026, responses will be required when submitting new apps or updates to the App Store.”
  • June 8, 2026: guidelines revised: “Introduction: revised kid and teen safety guidance. 1.2: new paragraph clarifies developer responsibilities for content that violates this guideline. 4.3(a) … 4.3(b): clarifies the basis for the guideline and adds examples.” The new 1.2 paragraph: “It is your responsibility to remove content that violates this guideline, your terms of service, or your community standards. If we find such content, we will ask you to remove it, and provide a plan to improve your compliance with this guideline.”2
  • February 24, 2026: “Apple will block users in Australia, Brazil, and Singapore from downloading apps rated 18+ unless they have been confirmed to be adults.”
  • February 6, 2026: guidelines revised “to clarify that apps with random or anonymous chat are subject to the 1.2 User-Generated Content guideline.”
  • February 3, 2026: “Starting April 28, 2026, apps and games uploaded to App Store Connect need to meet the following minimum requirements: iOS and iPadOS apps must be built with the iOS 26 & iPadOS 26 SDK or later.”
  • November 13, 2025: a new 4.1(c), “you cannot use another developer’s icon, brand, or product name in your app’s icon or name, without approval from the developer”; 5.1.2(i) now says you must “clearly disclose where personal data will be shared with third parties, including with third-party AI, and obtain explicit permission before doing so”; 2.5.10’s ban on “empty ad banners or test advertisements” was deleted; a new 1.2.1(a) requires creator apps to use “an age restriction mechanism based on verified or declared age.”
  • October 29, 2025: the submission model widens: “you can now send additional items to App Review independent of an existing submission,” including Game Center features.
  • July 24, 2025: “The updated age rating system adds 13+, 16+, and 18+ to the existing 4+ and 9+ ratings,” with answers required “by January 31, 2026.”

The submission model matters more than it sounds. App Store Connect Help: “The first consumable, non-consumable, auto-renewable subscription, and non-renewing subscription In-App Purchase of each type must be submitted with a new app version,” and “The app version, subscription group, and every subscription or In-App Purchase you want reviewed together must all be added to the same draft submission before you click Submit for Review.” For Game Center: “Include components in the same submission as your app version if this is the first time you’re adding them to an app.”12 Adding the version to a draft and submitting it, which is what the page’s Add for Review and Submit for Review buttons do, leaves products you didn’t add in “Ready to Submit” and Game Center components in “Prepare for Submission,” unreviewed (those are App Store Connect’s own status labels, read off our app’s records).24 The 2.1 rejection about products “not … submitted for review” is this model biting. It bites once: after the first item of a type is approved, the same help says, “you can submit additional items of that type without a new app version,” and “For subsequent components, you can submit them without an app version.”12

A checklist from the rules and the record

The sections above, as a list to run against your own listing the night before. The first seven are things rejection messages show reviewers checking; the last five are what the rules require whether or not anyone has posted a rejection for them.

  1. The app launches and whatever a reviewer can reach works, on an iPhone and on an iPad, on the current OS, without a crash, a blank view or placeholder text (2.1, 2.4.1).
  2. Anything behind a login has a demo account in App Review Information (Apple’s help: it “must not expire”), or a demo mode agreed in advance (2.1(a)).10
  3. Every in-app purchase and subscription you want reviewed is in a submission, and the first of each type is in the same one as a version, visible in the app and buyable in the sandbox; a restore mechanism exists for anything restorable (2.1(b), 3.1.1).
  4. The subscription screen shows the name, the length, the amount to be billed as the most prominent pricing element, and links to the privacy policy and Terms of Use; the listing’s description links the Terms of Use if you use Apple’s standard EULA (3.1.2, Schedule 2 §3.8(b)).2125
  5. The support URL opens a page with a way to contact you; the privacy policy URL opens the policy; both are also reachable inside the app (1.5, 5.1.1(i)).
  6. Every permission prompt’s purpose string says what the data is for, with an example (5.1.1(ii)).
  7. Every background mode you declare does what it says while the app is in the background (2.5.4); a framework linked for a feature the app doesn’t have has drawn a rejection too.
  8. The privacy label matches what leaves the device and stays anywhere; the age-rating answers match the content, including the September 2026 social-media questions (5.1.2, 2.3.6).
  9. Screenshots show the app in use on the right device; the name, subtitle, screenshots and previews carry no prices or terms; keywords carry no other apps’ or companies’ names and no trademarked terms you aren’t authorized to use (2.3.3, 2.3.7, and the keyword field’s own rules).26
  10. If the app shows what its users make or lets them meet (1.2’s “user-generated content or social networking services”), there is filtering, a report path with a response, a block, and published contact information, and nothing in it is a service “used primarily for … random or anonymous chat” (1.2).
  11. Third-party content in the app has a terms-of-use basis you could show on request (5.2.1, 5.2.2).
  12. The review notes describe every non-obvious feature “with specificity” (Apple’s words: “generic descriptions will be rejected”), name where the purchases are, explain anything the reviewer cannot exercise, and say that backends are live (2.3.1(a), 2.1(b), “Before You Submit”).2

Red-teaming our own app against it

Kiradex is a free collector’s app for Pokémon trading cards: every set, have and want lists, a camera scanner, a 3D card viewer, prices from the public TCGdex catalog, a small “plaza” where collectors’ pixel avatars share a square through our relay server, Game Center achievements, six “gift” tips and one Patron subscription in two lengths, all of which unlock cosmetic flair. Its 1.0 went into App Store Connect on October 4 from Xcode 27 (the Duo post has why), and on October 5 I ran the list above against it. Here is what that found, with the fix, grouped by the checklist item it falls under.24

Item What the reviewer would have found What we did
4. Terms of Use link in metadata The description had neither the Terms of Use nor the privacy link. This is the form letter quoted above. Added both lines to the description; App Store Connect’s API accepted the edit while the version was Waiting for Review.
3. Products in the submission The first submission carried the app version only. All eight products sat in “Ready to Submit.” Cancelled, resubmitted with the six gift versions, both subscription versions and the group version as items.
3. Game Center in the submission The Game Center page showed all fourteen achievements and the leaderboard as “Prepare for Submission.” A night-before attempt to “release” them through the API turned out to block the review: the page’s own request to add one came back 409, “You have non-live Achievement releases. Please delete or release them.” Deleted the fifteen releases, added the achievement and leaderboard versions to a draft submission, submitted. Twenty-five items in one submission.
3. Restoring tips Six consumable tips that add up to permanent flair is the pattern a reviewer can read as products “that function as non-consumable products” without restore; a developer posted that exact 3.1.1 message in September 2025.27 The app reads lifetime giving from the App Store’s own transaction history each time (Transaction.all, with SKIncludeConsumableInAppPurchaseHistory in Info.plist, which Apple documents as putting “all finished consumables” back into that history28), so nothing is kept on the device to lose, and a Restore Purchases button on the Patron screen calls AppStore.sync(). The app’s own walk buys a gift in Xcode’s local store and sees the tier on the card and the Patron screen; a reinstall or a second device has not been walked. The review notes now say how it restores.
9. Prices in product names The gifts were named “Gift $1” through “Gift $100,” with descriptions like “A dollar toward keeping Kiradex free.” Wrong outside the US (a “$1” gift is €0.99), and against the spirit of 2.3.7’s rule that app names, subtitles, screenshots and previews “should not include prices,” even though the rule does not name product names and I found no rejection for it. Renamed without a currency (Small, Gift, Generous, Big, Grand, Legendary) and reworded; those fields are locked in review, so this cost a cancel and resubmit.
9. Trademark in keywords “pokemon” led the keyword list. App Store Connect’s own field help: “Names of other apps or companies aren’t allowed.”26 Keywords aren’t public, so this is a rule, not a thing reviewers have been seen to catch. Removed. The word stays in the description, where it describes what the cards are, next to the non-affiliation line.
12. Review notes Specific, but silent on the things a reviewer would ask about: that nobody can type in the plaza, how tips restore, why the privacy label says nothing is collected. Rewritten: the four 1.2 mechanisms (fixed lines, Report, Hide, the support email), the restore path above, the scanner’s no-card state, the plaza server’s health URL, that an empty plaza is normal.
8. Age rating Answered “No” to messaging and “None” to contests. The plaza lets people choose lines to say to each other; the Collector Level board is, by Apple’s definition, “Events that allow users to compete with one another for rankings.”29 Changed to messaging “Yes” and contests “Infrequent.” The computed rating stayed 9+.
8. Privacy label “Data Not Collected” rests on the relay keeping nothing, and Apple’s definition turns on retention. The relay’s code holds its rooms in memory, writes no store, and its own log names nobody (a report is logged by its reason and its room). But the web server underneath it wrote each connection’s client address, passed through from the host’s proxy, into the host’s log, where it stays as long as the host keeps logs. Turned that connection log off: two flags, a redeploy, and a read of the host’s log afterwards, which showed a connection as a room name and a head count on arriving and leaving and nothing else; a report, by the code, logs a reason and a room. What the old log already holds stays until the host’s retention runs out, and the host’s own edge, which forwarded the address, keeps whatever connection logs it keeps under its own policy. Apple’s definition turns on what we or our partners can access beyond servicing the request, and its page says nothing about a hosting provider’s own request logs; “Data Not Collected” is our reading of that, and the privacy policy now says the host may keep connection logs with an address for a limited time.
5. Privacy policy in the app Linked on the Patron screen, two taps into Me, and nowhere else. 5.1.1(i) asks for “easily accessible.” A Privacy section in Settings with both links (next build).
1. Placeholder content The newest set in the catalog has no card pictures and no logo, so the first tile on the Dex was an empty dark window with a placeholder glyph. 2.1(a) says “other temporary content should be scrubbed.” The window now shows the set’s name when the catalog has nothing else (next build).
1. iPad Never run on one. Walked on an iPad Air (M3) simulator on iOS 27.0, the device named in the May 2026 rejection: the app runs in the iPhone-compatibility window with its bars, tiles and status screen in place, the opening plays, a set opens, a card stands on its 3D stage with its buttons, the collection lists it. The walk’s one failure on the first pass was its own: the new set-name fallback gave the test two “Kira Base” texts to choose from.
7. Entitlements and manifest Checked what Apple recorded for the build: push and iCloud entitlements set to production, the privacy manifest declaring UserDefaults, no required-reason API undeclared, the background task registered. Nothing to change.

Three things stay open and are judgment, not checklist. The card images are The Pokémon Company’s; TCGdex’s terms say “Pokémon card information and images are licensed by The Pokémon Company,” and I read nothing in them that grants rights in the images.27 Independent card apps with “Pokémon” in their names or subtitles and card images are live and were updated on the store in September 2026, one of them with the same kind of disclaimer we carry, which says Apple tolerates the category today, not that a licence exists.27 And the plaza, where 1.2 asks two different things. Apps with “user-generated content or social networking services” must have four mechanisms, and it has a version of each (fixed lines for filtering, a report path, a hide, a published address), said in the notes. Separately, services “used primarily for … random or anonymous chat” “do not belong on the App Store,” and February’s revision put random or anonymous chat squarely under the rule; strangers with generated handles saying fixed lines to each other is near that line. What we can say is that the plaza is one tab of a collector’s app, nothing in it is typed, and the lines are ours; whether a reviewer reads a square of pixel collectors as a chat service is theirs to decide. The third is smaller: 3.1.2(a) says a subscription “must provide ongoing value to the customer,” and Patron’s value is the standing each month adds on the ladder and the flair that standing unlocks, which is cosmetic; whether that is ongoing value is the reviewer’s reading too.

One more, which is why the build itself was not settled when this was written: on the simulators, the full-screen card viewer has come up black on a second or third opening in a share of the regression walk’s runs (about half of them, by the walk’s record as of October 4), with the 3D scene whole and ticking underneath, on two iOS runtimes.24 Making the cover opaque made the first opening black on every try, which pointed at a texture made for the inspector’s RealityView and reused by the viewer’s; each stage now makes its own. After that, the walk’s five runs on October 5 split: one had all three openings lit, the second at eight seconds rather than four; four had a later opening black through sixteen seconds (in one of them the second opening lit at twelve), and in one of those, frames of the simulator’s own display taken from the Mac were black too, with the Close button drawn, so it is the stage that draws nothing, not the test’s screenshot. Waiting eight seconds between openings instead of one changed nothing, and making the card’s mesh per stage, as the texture now is, changed nothing either. The readings look like a delay with a long tail rather than a fixed black, though a handful of runs cannot tell those apart, and the cause is not in the logs, which read the same for a lit opening and a black one. Whether a real iPhone does it at all is the question that matters; that device walk had not run when this was published, and the build in review carried the question. Later the same morning: the phone answered. The same walk, same build, ran on an iPhone 17 Pro Max on iOS 27.0 over the local network, and all three openings lit at four seconds, each reading 0.55 of the stage lit, the same figure as the simulator’s good openings.24 The black is the simulator’s renderer; on hardware the viewer draws first time, every time, and guideline 2.1 has nothing to catch.

Key Takeaways

  • The guideline Apple itself points at first is 2.1, completeness: crashes, blank screens, missing demo access, products that are not in the submission. Apple’s figure is “over 40% of unresolved issues”; in the last week it ranked, 2.1 was 42 percent of rejections.16
  • Assume an iPad. Reviewers have run iPhone-only apps on one for years, and an app being “intended primarily for use on iPhone” is not a defence Apple accepted.7
  • The submission is a container. For a new app, put the version, every product, the subscription group and the Game Center components in it; submitting the version alone leaves the rest unreviewed. The first product of each type, and the first Game Center components, need a version with them; once a type has been approved and the app has an approved version, later products of that type and later components can go on their own.12
  • Templates tell you the check. The EULA line, the support-URL line, the purpose-string line: each names the field the reviewer opened. Read them before you submit, not after.
  • Do the pass yourself. Ours was a listing three people had read, and it still had eleven things to fix, most of them under guidelines in the top-ten list above (2.1, 5.1.1, 3.1.1, 2.3.7).

FAQ

How long does App Review take in 2026?

Apple’s published figure, in two wordings: “On average, 90% of submissions are reviewed in less than 24 hours,” and “We review over 90% of apps within 24 hours,” undated and without a method.15 The 2021 trial added two more, from testimony: a service-level target of 50 percent within 24 hours and 90 percent within 48, and Apple’s claim that it was completing 96 percent within 24 hours.13 The guidelines add that complex apps “may require greater scrutiny,” that repeat violations mean review “will take longer to complete,” and that after approval “it can take up to 24-hours for your app to appear on all selected storefronts.”12

How many apps does Apple reject?

In 2025, 2,093,244 of 9,100,620 submissions, by the transparency report; “over 2 million” of “more than 9.1 million” in the Newsroom’s words. That is 23 percent of submissions. Apple does not publish a rate per app.34

Does App Review test on an iPad if my app is iPhone-only?

Often enough that you should build for it. Rejection messages have named the iPad and its OS for iPhone apps since 2017, and Apple has told a developer that iPhone apps “must still be formatted correctly and behave properly when running on iPad.” Whether every review uses one, Apple doesn’t say.7

If your app has auto-renewable subscriptions and you use Apple’s standard EULA, the rejection template says so in as many words: “include a link to the Terms of Use in your App Description.” Apple’s subscriptions page: “your app and App Store metadata must include links to your Terms of Use and Privacy Policy.”825

Are in-app purchases reviewed with the app automatically?

No. They are items you add to a submission; the first of each type “must be submitted with a new app version,” and once that type has been approved, later ones can go “without a new app version, as long as your app has at least one approved app version at the time of submission.” Game Center achievements and leaderboards work the same way the first time.12

What changed for new apps in 2026?

Social-media answers in the age-rating questionnaire are required since September 2026; uploads must use the iOS 26 SDK or later since April 28, 2026; random or anonymous chat is under guideline 1.2 since February; and the guidelines’ June 8 revision clarified the duty to remove violating content and to provide a plan when Apple asks for one.9

Related on this site: the iPhone Duo preparation post is where this app’s submission calendar is worked out; the screenshots post covers the 2.3.3 and 2.3.7 rules for the product page; the social-media declaration post covers the new age-rating questions; and the privacy manifest deep dive covers the privacy manifest.

Sources


  1. Apple, App Review, Apple Developer (the older URL /app-store/review/ redirects here), fetched October 5, 2026, quoted: the review-time sentence, the 2.1 share of unresolved issues, the “Avoiding common issues” list, and the bug-fix offer. The page carries no date. ↩↩↩↩↩↩↩↩↩

  2. Apple, App Store Review Guidelines, fetched October 5, 2026; the page prints “Last Updated: June 8, 2026.” Quoted: the “After You Submit” section and guidelines 1.2, 2.1, 2.3.1, 2.3.6, 2.3.7, 2.4.1, 5.1.1. ↩↩↩↩↩↩↩↩↩↩↩

  3. Apple Newsroom, The App Store stopped over $2.2 billion in potentially fraudulent transactions in 2025, May 20, 2026; The App Store prevented more than $9 billion in fraudulent transactions, May 27, 2025; App Store stopped over $7 billion in potentially fraudulent transactions in four years, May 14, 2024. All fetched October 5, 2026; figures and sentences quoted as printed. ↩↩↩↩↩↩↩↩↩↩

  4. Apple, App Store Transparency Report, the 2025 report and the 2023 and 2024 PDFs linked from the index, fetched October 5, 2026. Counts quoted as printed: 2023 6,892,500 reviewed / 1,763,812 rejected (Performance 1,132,613, Legal 420,914, Design 295,497, Business 192,884, Safety 103,629); 2024 7,771,599 / 1,931,400 (1,235,471; 445,696; 378,300; 209,845; 116,105); 2025 9,100,620 / 2,093,244 (1,354,418; 495,673; 415,532; 283,820; 151,159), with 387,087 “approved after rejection.” Apple’s footnotes: “Apps may be submitted to App Review multiple times before they are approved for release on the App Store” and “App submissions may be rejected for noncompliance with one or more guidelines.” The percentages in the text are the author’s arithmetic on those counts. ↩↩↩↩↩↩↩↩↩↩↩

  5. Apple, App Store, apple.com, fetched October 5, 2026 (“Every week, nearly 500 dedicated experts around the world review over 130K apps. In 2024, more than 1.9M app submissions were rejected for reasons that include privacy violations and fraudulent activity.”), with Internet Archive captures of June 1, 2024 (“over 500 dedicated experts … over 100K apps”) and October 1, 2025 (the current wording); Apple, Developing for the App Store, fetched October 5, 2026 (“We review over 90% of apps within 24 hours.”); Apple, “App Store — Principles and Practices,” 2019, read through the Internet Archive’s May 29, 2019 capture on October 5, 2026 (the live URL now redirects to the App Store page), quoted as printed. ↩↩↩↩↩↩↩↩↩

  6. Apple, “Common App Rejections,” the retired page at developer.apple.com/app-store/review/rejections/, as captured by the Internet Archive on December 2, 2018, fetched October 5, 2026: “Top 10 reasons for app rejections” for the “7-day period ending April 3, 2018,” with “Total Percent of App Rejections 88% Top 10 Reasons 12% Other Reasons (<2% each).” The live URL now redirects to the App Review page. Captures of June 4, 2016 and May 29, 2015 list the weeks ending April 24, 2016 and April 20, 2015 under the old numbering. ↩↩↩

  7. Developers’ posts quoting App Review, fetched October 5, 2026: Stack Overflow 70815055 (January 2022; the crash log’s "modelCode" : "iPad7,5"); 41730125 (January 2017; “did not run at iPhone resolution when reviewed on iPad running iOS 10.2”); 73578508 (September 2022; the “formatted correctly and behave properly when running on iPad” and “unable to provide further information regarding our review environment” replies); 79936983 (May 2026; the developer’s own description of the reviewer’s report, “on an iPad Air 11-inch (M3) running iPadOS 26.4.2”). These are developers’ transcriptions of Apple’s messages, not Apple publications. ↩↩↩↩

  8. Developers’ posts quoting App Review’s messages, fetched October 5, 2026, each a transcription by the developer who received it: subscriptions, Stack Overflow 74278461 (November 2022; an answer dated September 29, 2025 quotes the same “unable to find the following required item(s) in your app’s metadata” line) and 76209970 (May 2023, the binary-and-metadata variant); crash on launch, 70815055 (2022) and 33346091 (2015, the same sentence under the old numbering); demo account, 43356595 (2017); in-app purchases not submitted, 79002395 (September 2024); purpose string, 64691575 (2020); support URL, 43278066 and 59207086 (2017, 2019); screenshots, 53106409 (2018); app name, 43241848 (2017); minimum functionality, 56446210 (2019); spam, 45667592 (2017); user-generated content, 43497308 (2017); background audio, 75008240 (2023); PassKit, 76209970 (2023); intellectual property, 65591763 (2021, a 5.2.3 case). Apple may have reworded any of these since; only the subscription and iPad messages are confirmed in use after 2024. ↩↩↩↩↩↩↩↩↩↩↩↩

  9. Apple Developer News, items dated from the feed’s pubDate, fetched October 5, 2026: August 5, 2026 (creative assets); July 9, 2026 (social-media questions); June 8, 2026 (guidelines revised); February 24, 2026 (age requirements by region); February 6, 2026 (random or anonymous chat under 1.2); February 3, 2026 (SDK minimums); November 13, 2025 (guidelines revised); October 29, 2025 (submission items); July 24, 2025 (age ratings). Quoted as printed. ↩↩↩

  10. Apple, App Store Connect Help, Overview of submitting for review (“Submissions may not be reviewed in the order you submit them”; “Each platform can have one app version submission under review at a time”) and Platform version information (the App Review Information fields: the demo account “must not expire”; the Notes field “can contain up to 4000 bytes”), fetched October 5, 2026, quoted. ↩↩↩↩

  11. Apple, App Store Connect Help, App and submission statuses, Reply to App Review messages, and Manage a submission with unresolved issues, fetched October 5, 2026, quoted. The term “Resolution Center” no longer appears on these pages; the current name is the App Review section. ↩↩↩↩

  12. Apple, App Store Connect Help, Submit an In-App Purchase, Overview for configuring in-app purchases, Submit an app (Add for Review adds the version to a draft; “Your submission won’t be sent to App Review until you submit it for review in the next step”), and Submit Game Center components, fetched October 5, 2026, quoted. ↩↩↩↩↩

  13. Epic Games, Inc. v. Apple Inc., No. 4:20-cv-05640-YGR (N.D. Cal.), Rule 52 Order After Trial on the Merits, Dkt. 812, September 10, 2021 (Judge Yvonne Gonzalez Rogers), RECAP copy, fetched October 5, 2026. Quoted: “only 500 human reviewers” (printed p. 102, footnote 501 citing Kosmynka’s trial testimony); the Service Level Agreement and the 96 percent claim (pp. 36–37, footnote 216 citing Kosmynka and Federighi); what the human reviewer does (p. 106); the error rate “around 15% in 2015” (p. 107, footnote 525 citing Shoemaker’s deposition; the court adds that a rejection-rate statistic “says nothing about the error rate”). ↩↩↩↩↩↩

  14. Declaration of Trystan Kosmynka, Senior Director of App Review, Apple Inc., in Epic v. Apple, Dkt. 821-10, executed October 8, 2021, RECAP copy, fetched October 5, 2026; ¶1 (the Executive Review Board “creates policies”), ¶2 (static and dynamic analysis plus human review), ¶6 (the submission and rejection figures), quoted as written, including “submissions of apps and app reviews per week.” ↩↩↩↩

  15. Kif Leswing, Inside Apple’s team that greenlights iPhone apps for the App Store, CNBC, June 21, 2019, fetched October 5, 2026. Its sources are former reviewers and “people familiar with the process,” unnamed; the headcount, languages, quota, App Claim, iPad, SLA, calls-per-week and board descriptions are quoted from it, with “Apple says” figures attributed to Apple through CNBC. ↩↩↩↩↩↩

  16. Epic Games’ Proposed Findings of Fact and Conclusions of Law, Epic v. Apple, Dkt. 407, April 8, 2021 (redacted public version), RECAP copy, fetched October 5, 2026: ¶521(a) (13 and 6 minutes, citing Shoemaker Dep. 56:15–17), ¶521(b) (50 to 100 apps per day, citing PX06), ¶521(c) (“less than a minute”), ¶519 (“review[s] … not test[s],” citing PX140), ¶529 and ¶529(a) (the Fraud Engineering Algorithms and Risk team and its head, citing Friedman Dep. 82:22–83:3, 94:9–19 and exhibits PX250–251). These are Epic’s characterizations of deposition testimony and of exhibits, selected by Apple’s opponent, not the transcripts or documents themselves; the court called Shoemaker’s excerpts “limited.” ↩↩↩↩↩

  17. Mark Gurman, Inside the Apple Team That Decides Which Apps Get on iPhones, Bloomberg, May 28, 2019, read through the Internet Archive’s May 29, 2019 capture on October 5, 2026 (the live page sits behind a bot wall); Neil Long, Apple’s Phil Schiller refuses to fix app review…, MobileGamer.biz, June 22, 2023 (“four hours in a room arguing about apps”); Phillip Shoemaker, I built Apple’s App Review team. The approval you trust is a snapshot, not a promise., LinkedIn, September 14, 2026, a first-person post promoting his book. Shoemaker ran App Review from 2009 to 2016 and is a former employee with stated grievances; the words in quotation marks are his own, the three-reviewers, one-set-of-eyes and 30-to-100 details are Bloomberg’s paraphrase of his remarks, and his numbers are unverified. ↩↩↩↩

  18. Adi Robertson, Epic needs to tarnish Apple’s App Store halo, The Verge, May 6, 2021 (Kosmynka’s “It’s a human process. We do make mistakes”); Filipe Espósito, App Store review process has over 500 human experts; less than 1% of rejections are appealed, 9to5Mac, May 7, 2021, a press summary of Kosmynka’s written trial testimony; Elizabeth Lopatto, In Epic v Apple, everybody is losing at the game of defining games, The Verge, May 7, 2021 (“only 500 people actually do the ‘human review’”). All fetched October 5, 2026. ↩↩↩↩↩

  19. Apple job postings read on October 5, 2026: App Review Specialist, Spanish Language (posted September 23, 2026, Sunnyvale; minimum qualifications “Fluency in English and Spanish” and “Able to work Tuesday through Saturday, 40 hours per week”), the French, Chinese and Turkish roles (the latter two “Sunday through Thursday, 40 hours per week”), and Tools and Automations Lead, App Review (posted July 2, 2026, Cupertino), quoted as printed. ↩↩

  20. Apple, Validating receipts with the App Store, StoreKit documentation, fetched October 5, 2026, quoted. ↩

  21. Apple, Apple Developer Program License Agreement, Schedule 2, section 3.8(b), fetched October 5, 2026, quoted. ↩↩

  22. Apple, App privacy details on the App Store, fetched October 5, 2026, quoted; the definition of “collect” and the “immediately discarded” example. ↩

  23. RevenueCat, The ultimate guide to App Store rejections (June 2023, updated June 6, 2024); Runway, Everything you need to know about submitting to the App Store (and avoiding rejections) (undated); Bitrise, App Store review time (2022, modified April 2024). Vendor sources, fetched October 5, 2026; none states a dataset. The $!{infoPlistKey} template text appears on Apple’s App Review page and in Runway’s article. The search for datasets was by direct fetch, without a web search engine, so a study may have been missed. ↩

  24. Author’s work on Kiradex, October 4–5, 2026: App Store Connect’s records read through its API (version, products, submissions, Game Center components, age-rating declaration, build entitlements) before and after each change; the 409 response body quoted from the App Store Connect page’s own request when adding an achievement (“You have non-live Achievement releases. Please delete or release them”); the relay’s source read for what it logs and stores (server/app/main.py and rooms.py: rooms in memory, no store, its own log lines a room name and a count, a report its reason and room), and uvicorn 0.37’s WebSocket handler, which logs the client address on accept at INFO through its uvicorn.error logger, silenced with --no-access-log --log-level warning, deployed to the relay’s host on October 5 at 16:42 UTC and read back from the host’s log after a test connection (the log held the container’s start line, the two room lines, and nothing else); the viewer walk testTheViewerShowsTheCardEveryTime run on the iOS 27.0 and 27.2 simulator runtimes since October 4 (the repo’s roadmap recorded it failing “about half its runs” on the evening of October 4; the opaque-cover experiment and the texture finding are in the repo’s commit 1800ceb), and its three October 5 runs after that change as printed by the test: “viewer opening 2: 0.0 lit at 4 s, 0.0 lit at 8 s, 0.55078125 lit at 12 s” and “viewer opening 3: 0.0 lit at 4 s, 0.0 lit at 8 s, 0.0 lit at 12 s, 0.0 lit at 16 s”; then “viewer opening 2: 0.0 lit at 4 s, 0.546875 lit at 8 s” with openings 1 and 3 lit at 4 s; then “viewer opening 2: 0.0 lit at 4 s, 0.0 lit at 8 s, 0.0 lit at 12 s, 0.0 lit at 16 s” while frames of the simulator’s display taken from the Mac every two seconds (xcrun simctl io … screenshot) read a mean brightness of 0.0 through the same window and 134–163 during the first opening, with the Close button drawn in both; and the same walk on a physical iPhone 17 Pro Max (iOS 27.0, Xcode 27.0, connected over the local network), later on October 5: “viewer opening 1: 0.546875 lit at 4 s”, “viewer opening 2: 0.546875 lit at 4 s”, “viewer opening 3: 0.546875 lit at 4 s”, passed. ↩↩↩↩

  25. Apple, Auto-renewable subscriptions, App Store, fetched October 5, 2026: “Please note that your app and App Store metadata must include links to your Terms of Use and Privacy Policy,” the sign-up screen requirements, and “the amount that will be billed must be the most prominent pricing element in the layout”; and Provide a custom license agreement, App Store Connect Help: “If you don’t provide a custom EULA, the standard EULA is applied to your app and the license agreement link isn’t shown on your App Store product page.” ↩↩

  26. Apple, App Store Connect Help, Platform version information, the Keywords field (“Names of other apps or companies aren’t allowed”), and Creating your product page, keywords to avoid (“Unauthorized use of trademarked terms, celebrity names, and other protected words and phrases”), fetched October 5, 2026. ↩↩

  27. TCGdex, Terms and Conditions (“Last Updated: 2024-07-15”): “Pokémon card information and images are licensed by The Pokémon Company,” and that the site’s content is CC BY-SA 4.0, fetched October 5, 2026; the App Store listings of Dex - for TCG Collectors (subtitle “Track & Scan Pokémon Cards,” with a non-affiliation disclaimer; last version September 16, 2026), Rare Candy: Scan Pokémon Cards (September 23, 2026, no disclaimer found), Pokellector: Card Collector (subtitle “Modern Pokemon TCG Collecting”; October 17, 2024) and Collectr - TCG Collector App (September 30, 2026, no disclaimer found), read through Apple’s iTunes Lookup API on October 5, 2026; and an Apple Developer Forums thread of September 2025 (800637, an anonymous developer) quoting a 3.1.1 rejection: “Your app uses intermediary currency to purchase items that function as non-consumable products but still does not include a restore mechanism.” ↩↩↩

  28. Apple, SKIncludeConsumableInAppPurchaseHistory, Information Property List, Bundle Resources documentation, fetched October 5, 2026: “A Boolean value that determines whether StoreKit includes finished consumable In-App Purchases in transaction information”; when it is set, “StoreKit includes all In-App Purchase transactions — including all finished consumables — in the transaction information” from Transaction.all and the other APIs the page lists; available from iOS 18. The page asks that you “be sure you have a way to reconcile a customer’s consumable transactions on your server, not only on the device,” to avoid delivering content twice after a reinstall; Kiradex delivers nothing twice, because the tier is a sum of the history, not a grant. ↩

  29. Apple, App Store Connect Help, Age ratings values and definitions, fetched October 5, 2026: “Contests: Events that allow users to compete with one another for rankings, rewards, or the achievement of personal goals,” and “Messaging and Chat: Users can directly communicate with one another through features within the app,” with 9+ covering “Infrequent cartoon or fantasy violence.” ↩

Artículos relacionados

Prepare Your App for iPhone Duo: A Worked Example

Prepare and submit an app for iPhone Duo: the SDK stamp that switches the layout on, a real app walked through every pos…

64 min de lectura

The App Store's Social Media Checkbox and What It Costs

September's App Store rule makes every submission declare social media capabilities. The under-13 carve-out costs an ent…

30 min de lectura

A Pixel Town Alive When Nobody Is Online: NPCs and Ghosts

Emerald's 158 townspeople measured, Stardew's schedules, other players' traces, presence buffers and Apple's child rules…

104 min de lectura