Apps built with Lovable, Bolt, Replit, Base44 or Cursor can go into the App Store and Google Play, but most of them are web apps, so they need packaging as a native app or a native front end first. Then they need what every store app needs: developer accounts, a privacy policy, account deletion, a demo login and a security check. There are three realistic paths: wrap and harden, build a companion app on the same backend, or rebuild.
These tools are a real step forward. You've gone from an idea to something people can click through, often with a working database behind it. That's the hardest part of a brief, done. What's left is the part the stores care about. Some of these tools now have a publish button too, so for many owners the question has moved from "how do I get it into the store?" to "will it pass review, and is it safe to run once it's live?"
What these tools usually produce
Most AI app builders produce a web app: a front end that runs in the browser, often connected to a hosted database such as Supabase. Lovable apps, for example, are web apps built on React. Some tools now offer a mobile project option, often built on Expo and React Native, and Cursor is a code editor, so the output is whatever you asked it for. Check what you actually have before you choose a path.
That matters because Apple's guideline 4.2 asks for "features, content, and UI that elevate it beyond a repackaged website." A web app placed in a wrapper with nothing added is the classic rejection.
If your tool has a publish button
Two of these tools now take you most of the way into the stores themselves. From their own documentation, as read on 2026-11-11:
| Tool | What its store publishing does | What it doesn't do yet |
|---|---|---|
| Replit | Builds a native React Native app with Expo; its Launch feature "creates that build, signs it, and uploads it" to App Store Connect from the Publishing tab | "Publishing to Google Play is not yet supported by Replit" |
| Base44 | Runs "your published Base44 app inside a secure web view" and generates the files to submit to both stores | Base44 apps "don't support Apple In-App Purchase or Google Play Billing yet" |
Sources: Replit: upload your app to App Store Connect, Replit: native mobile apps, Base44: submitting your app to app stores.
Both use your own Apple and Google developer accounts, and neither button passes review for you. A Base44 app is a wrapper, so Apple's 4.2 test applies in full. Base44 also says most content and design changes "also appear in your app without sending a new version" to the stores. That's handy, but those changes reach people's phones without review, and Apple's new guideline 2.5.19, added on October 16, 2026, says "Don't include any hidden, dormant, or undocumented features in your app". Ship new features as a reviewed update.
What's missing for the stores
| Area | What the stores or users need | Common gap in AI-built apps |
|---|---|---|
| Native build | An iOS and Android build, signed and uploaded | The app only runs in a browser, or the tool's publish flow covers one store |
| Accounts | Apple and Google developer accounts in your name | Not set up, or tied to a tool's own account |
| Privacy | Privacy policy, Apple privacy labels, Google Data safety form | No policy, or labels that don't match the app |
| Account deletion | In-app deletion, plus a web link for Google | Sign-up exists, deletion doesn't |
| Payments | Apple's in-app purchase rules for digital features | Card checkout that Apple won't accept for unlocks |
| Security | Database rules, hidden keys, safe sign-in | Keys in front-end code, tables anyone can read |
| Reliability | Backups, separate test and live versions | One database, no backups, changes go straight to live |
Each row is in the official rules. Apple's 5.1.1(v): "If your app supports account creation, you must also offer account deletion within the app." Google requires "a web link resource where users can request app account deletion" (Play Console help). On payments, Apple's guideline 3.1.1 says that to unlock features or functionality, such as subscriptions, "you must use in-app purchase," with an exception for link-outs in US storefront apps (guidelines).
The three realistic paths
1. Wrap and harden
Keep your web app, package it as a native app, and fix what the stores and a security review need: account deletion, privacy pages, Sign in with Apple if you offer Google login, a demo account and a few real app features, such as push messages, so it isn't just the website.
If Base44's publish flow already wrapped the app, this is the harden half.
Best when: the app works well on a phone screen and you need it in the stores soon. PerceptiaAI's Store Launch ($1,190 or $1,990) covers this: packaging for both stores, launch-blocking fixes, a safety pass and the Apple and Google must-haves.
2. A companion app on the same backend
Keep your database and backend, and build native mobile screens on top. Your web app stays as it is for desktop and admin use. This is the safest route on 4.2 because the mobile app is genuinely native. PerceptiaAI's Companion App is $3,490 or $6,500.
3. Rebuild
If the code is hard to follow, the data model has grown awkward or the tool limits what you can do, rebuild it. What you made isn't wasted: it becomes the clearest spec a developer could ask for. A PerceptiaAI rebuild starts with an App Preview, at the prices in the table below.
| Path | PerceptiaAI price | Keeps your code | Store risk |
|---|---|---|---|
| Wrap and harden | $1,190 or $1,990 | Yes | Medium: depends on what's added |
| Companion app | $3,490 or $6,500 | Backend, yes | Low |
| Rebuild | $3,990 or $7,500 | No; it's the spec | Low |
For generic web apps, PWAs and no-code tools such as Bubble or Glide, see how to get a web app into the app stores.
How an engineer reviews code from these tools
We use AI to draft code too, and a named engineer reviews and tests every change. Reviewing someone else's AI-built app follows the same steps. An engineer who hasn't seen it before should:
- Build it from a clean copy. If it only runs inside the tool, that's the first job.
- Read the database access rules. Can a signed-in user read other users' data? Can someone who isn't signed in?
- Search for keys and secrets in the front-end code and in the repository's history.
- Walk every sign-in path: sign-up, password reset, social login, account deletion.
- Test payments end to end in the store or payment sandbox, including failed and refunded payments.
- Check what happens when things fail: no signal, an expired login, a slow server.
- Map every account to a named owner: hosting, database, domain, payments, stores.
- Confirm backups exist and that one restore has actually worked.
To run the security half yourself first, use our security checklist for vibe-coded apps. PerceptiaAI's App Check does this review for $249 or $449, in 2–4 working days. You get a plain-English report on what works, the safety problems, likely store rejections and who controls which accounts, with a clear fix, launch or rebuild answer and one fixed price.
Keep building after launch, safely
Many owners want to keep using the AI tool once the app is live, and that's reasonable. The rule is simple: every change gets checked before it reaches real users. That matters more when the tool can push changes straight into the installed app, as a web-view app does. App Care includes exactly that, "a check of any change you or an AI tool make, before it goes live," along with crash watching and store-rule updates. Plans start at $129 a month.
- Code in a repository you own, not only inside the tool
- Separate test and live versions
- Database rules that stop users reading each other's data
- No secret keys in front-end code
- In-app account deletion and a privacy policy
- In-app purchase for digital unlocks on iPhone, where Apple requires it
- Backups switched on and one restore tested
- A demo login ready for the reviewers
For how a takeover runs, step by step and with prices, see the concept app: Deskhaus, a rejected app rebuilt and relaunched. If yours was built in a no-code tool like Bubble, see Shiftbay, a Bubble app rebuilt in code with its data moved over.
Changelog
- 2026-11-11: Rechecked Apple's guidelines 4.2, 3.1.1 and 5.1.1(v) and Google's account deletion article; the quotes still match.
- 2026-11-11: Added a section on publish buttons: Replit's App Store upload through Expo (no Google Play yet) and Base44's web-view wrapper for both stores.
- 2026-11-11: Added Apple's new guideline 2.5.19 and what the 2026 enforcement of guideline 2.5.2 against vibe-coding apps does and doesn't cover.
