Work out whether you need an app at all, then take the cheapest route that gets you there
There are three ways to get an AI-built web app onto a phone, and they differ by an order of magnitude in cost.
- Typical timeline
- 1 week to 12 weeks, depending on the answer
- Pricing
- The decision review is free
What this is actually about
The parts that decide whether the project takes three weeks or three months, and which nobody mentions until they bite.
- 01
The three routes differ by 10x, not by 10%
A progressive web app is days. A Capacitor shell with real native plugins is weeks. A full native rebuild is months. They are all called "making it a mobile app" and they are not remotely the same project, so being sold the wrong one is expensive in a way that is invisible until the invoice.
- 02
Most products need fewer native capabilities than they think
The usual list is push notifications, offline, and a home-screen icon.
- 03
A store listing is a distribution decision
For consumer products, being in the store is where discovery happens.
- 04
Choosing badly is recoverable, but not cheaply
Starting with a progressive web app and moving to native later is a normal, healthy path.
The sentences we hear most, and what we do about each one
Paraphrases of things founders have actually said on scoping calls. Left column is the experience; right column is the work.
"My users keep asking for an app"
This is genuine feedback and it is usually not literal. Users ask for an app when the mobile web experience is slow, when they have to log in every time, or when they cannot find it again.
How we helpWe look at what they are actually complaining about before recommending a build.
"An agency quoted me a rebuild and I do not know if that is right"
The quote is for the most expensive of the three routes, the reasoning is not written down, and there is no way to evaluate it without already knowing the answer.
How we helpThe decision review is free and it ends in writing: which route, why, what it costs, and what you would give up by taking the cheaper one.
"I need push notifications"
The most common single requirement, and the one where the platforms differ most. Support and reliability vary by platform and by how the app was installed, which makes "can a web app do push" a question with an annoying, conditional answer.
How we helpWe test push on your actual target devices before recommending a route, because the answer depends on the platform and the install path rather than on a blog post.
"I want to be in the store for credibility"
A legitimate reason, and one nobody says out loud because it sounds like vanity.
How we helpIt is a real reason and we treat it as one.
Everything that has to be true before you submit
We work through every one of these before the first submission, because finding them one rejection at a time is where the weeks go.
Route A - Progressive web app
Days. Installable, offline-capable, push on Android. No store listing, weaker iOS push.
Route B - Capacitor shell with native plugins
Weeks. Your existing web app plus real native capability and a store listing.
Route C - Native rebuild
Months. React Native or Flutter. The right answer when the app is the product rather than a channel.
The deciding question
What must this do that a browser cannot? If the list is empty, route A. If it is short, route B. If the interface itself is the product, route C.
Distribution
Do your buyers find you by searching a store? Consumer, usually yes.
Who maintains it
Every route adds a release process. Route C adds two. Be honest about the team you will have.
Deliverables
Everything lands in accounts you own, with the certificates and keys in your name.
- A written recommendation: route, reasoning, cost, and the trade-off
- Push notification testing on your actual target devices
- A cost comparison across all three routes for your product
- If route A: installable progressive web app with offline and push
- If route B: Capacitor shell with the native capability review requires
- If route C: a scoped rebuild plan you can take to any agency
Phases
Scope is fixed after phase one. Every phase ends in something you keep.
- 013–5 days
Decision review
Free. We look at the app, test what a web app can actually do on your target devices, and ask what your users are really complaining about.
- 021–12 weeks
Build the chosen route
Scope and duration follow the decision. Route A is often finished before route C has been scoped.
- 0330 days
Measure and revisit
Install rate, retention, notification opt-in and crash rate. If route A turns out not to be enough, you now have evidence for route B or C rather than an opinion.
What can go wrong, and what stops it
Picking the cheap route and outgrowing it
That is a success, not a failure - you learned what was needed with real usage instead of guessing.
Push not working the way a blog post promised
It is tested on your real target devices during the free review, not assumed.
A Capacitor shell rejected as a repackaged website
Route B only ships with genuine native capability. Where the product has no reason to have any, route B is not recommended in the first place.
Before you book the call
For a great many B2B tools, yes, and it is dramatically cheaper.
Yes, when the app does something a website cannot. Capacitor is not the problem; a wrapper with nothing added is.
A written recommendation with the reasoning and a cost comparison across all three routes.
Related tracks
Apple App Store
Apple rejects wrapped web apps under guideline 4.2, and that is the single most common outcome for an app built in Base44, Replit or Lovable and dropped into a WebView.
Read the planReact Native rebuild
Lovable, Bolt, v0 and most Base44 front ends are React. That is the whole argument for React Native: the component model, the state handling, the validation rules and every line of your API layer come across largely intact.
Read the planGoogle Play
Play is easier to publish to than the App Store and harder to plan around, because a new personal developer account has to run a closed test with a dozen real testers for fourteen continuous days before it can even apply for production.
Read the plan
Want a straight answer on the route?
The review is free and it ends in writing, including the case for the cheaper option.