What you need before building an app is shorter than the checklist you were handed.
What do you need before building an app? Three things, and two of them are free. An idea narrow enough to describe in one sentence. A decision about where the app keeps its data. A paid developer account, but only at the point where someone other than you installs it. Everything else on the usual app planning checklist is optional: the wireframes, the spec, the company, the logo, the funding. Most of it is better decided after the first version runs on a phone. For the process that comes after the prerequisites, the app development guide walks the whole thing end to end.
The honest answer splits by reader. If the app is for you, a handful of friends, or a team you can name, you need almost nothing to start. If it is going into the App Store or onto Google Play for strangers, the stores add requirements that are outside your control, and those are the only items on this page with a queue in them.
See what you actually still needThe short version
Two free decisions, one paid account, and nothing else before the first build.
Nothing technical is a prerequisite any more. A tool that writes the project for you removes the part of this list that used to be the longest: no toolchain to install, no simulator to configure, no starter repo to pick. What it cannot remove are the two decisions that are about your app rather than about code, which is scope and data.
So the working version of what you need before you build an app is this. Write one sentence naming who opens it and what they do in the first ten seconds. Decide whether the data ever leaves the phone. Then budget 99 USD a year for Apple and 25 USD once for Google, but only at the point where other people install the thing. The first two take an hour. The third is the only one that involves waiting for somebody else.
The three things that are genuinely required
Start with the one that is not about software. You need an idea narrow enough that you can name the person who opens the app and the single thing they do first. Not a category, not a market: one person, one action. A place to log which plants I watered is buildable this afternoon. A social network for gardeners is not an app, it is a company, and it stalls because nobody can say what the first screen shows.
Second, decide where the data lives, because that changes more of the build than any other choice you will make. An app that keeps everything on the phone needs no accounts, no server and no story about storing other people's records. The moment two people have to see the same list, or one person has to see it on two devices, you have added sign-in, a database and a bill that repeats. Both answers are fine. Arriving at the second one by accident in week three is not.
Third, a developer account, and only at the point where someone other than you installs the app. This is the single item here that costs money and depends on somebody else approving you, so start it early if you already have a date in mind. Nothing about it blocks the building, which is why it sits third rather than first on a list that most checklists put in the opposite order.
Try it
Which of these actually applies to you
Tap the statements that are true. What is left below is the required half of the list.
0 USD of accounts in year one
- One sentence: who opens it, and what they do first
- No server yet: the data can stay on the phone
- Nothing from either store yet
Nothing on this list renews while the app stays on devices you own.
Where the planning week usually goes wrong
The two decisions above take an hour between them. A typical app planning checklist takes a week, and the week goes into artefacts nobody opens again: a feature list the first build contradicts, a competitor grid, a written spec produced before anyone has tapped anything. None of that is required. Some of it actively hurts, because a feature list on paper is much harder to cut than an idea still in your head.
What does help is deciding now which features are not in version one. That is the entire discipline behind mvp app development, and on a first app it usually resolves to one screen, one list, and no settings page. If you cannot cut the idea down to that, the build is not what is standing in your way.
You do not need a designer either, and you do not need to have chosen a colour. You do need to know what the first screen shows, because otherwise the tool or the developer picks for you and you will spend the review arguing about it. A few conventions from app design basics beat a mood board, and both beat a design system nobody is going to maintain.
And you do not need a company. Both stores register individuals under their own legal name, and the organisation route is the slower one, not the faster one, as the next section shows. Incorporating before you have a working app is a tax decision wearing the costume of a technical prerequisite.
The one prerequisite that costs money
Apple states the fee plainly: the Apple Developer Program is 99 USD per membership year, with local currency shown during enrollment. To enroll as an individual you need an Apple Account with two-factor authentication turned on, your legal name in the name fields, and to be the legal age of majority in your region. That is the entire list for one person building one app.
Organisations get a longer list. Apple asks organisations for a D-U-N-S number so it can verify the legal entity. It also wants a work email on the organisation's own domain, a public website on that domain, and an enroller with authority to sign agreements. If you are one person with an idea, enroll as an individual and move on. Nothing you will miss in month one is on the other side of that.
What the money buys is distribution. The account is how any build reaches App Store Connect, TestFlight and eventually the store, so it is not a formality you clear on launch day. Apple's TestFlight page puts the ceilings at up to 100 internal testers holding a role on your team and up to 10,000 external ones. External testing needs App Review approval before invitations go out. Read that as: the account is what lets you hand a build to your first tester, not just to your first customer.
Apple Developer Program, what enrollment requires and what it costs
What changes the moment strangers install it
Google Play charges a 25 USD one-time registration fee, and you must be at least 18 to sign up for a Play Console account. Google also verifies who you are, and its own help page warns that you may be asked for a valid government ID and a credit card, both under your legal name. Neither store lets you publish anonymously, which is the part first-time publishers are usually surprised by.
There is a newer requirement that catches exactly this reader. Google says personal developer accounts created after 13 November 2023 have to meet testing requirements before an app can be made available on Google Play. Its help pages put that bar at a minimum of 12 testers opted in continuously for at least 14 days. If the plan was to publish quietly and see what happens, the plan now needs twelve real people and two weeks in it. Find the testers while you build, not after.
Apple has a rule worth reading before you choose your tooling. Guideline 4.2.6 of the App Review Guidelines covers apps made from a commercialised template or an app generation service. Those get rejected when the submitter is not the provider of the app's content. So use whatever writes the code, then submit under your own account, in your own name, with content that is yours. The guidelines say nothing about how the code was produced, and anyone telling you Apple rejects AI built apps is describing a rule that does not exist.
Past the accounts, each store has its own submission list: icon sizes, screenshots per device class, a privacy questionnaire about what the app collects, an age rating, a support URL. None of that blocks building, which is why it is not on this page's list of app prerequisites, and all of it blocks shipping. It is long enough to be its own page, so keep what the stores will ask for open on submission day rather than trying to memorise it now.
Google Play Console Help, registration fee and account verification
What each item on the checklist actually costs you
| The item | Costs money | Waits on someone else | Needed before you build | Needed before strangers install it |
|---|---|---|---|---|
| One sentence describing the app | No | No | Yes | Yes |
| A decision about where data lives | No | No | Yes | Yes |
| Apple Developer Program, 99 USD a year | Yes | Yes | No | for iOS |
| Google Play registration, 25 USD once | Yes | Yes | No | for Android |
| Wireframes, a spec, a company, a logo | sometimes | No | No | No |
Starting the build once the list is this short
If the one sentence exists and the data question has an answer, the build can start today. The first version is almost always smaller than the plan that preceded it, and the fastest way to find that out is to have something running on a phone by the end of the day.
Newly is an AI app builder. You describe the app in plain English and it writes a real React Native and TypeScript project you own, runs it on cloud iOS and Android simulators while it builds, and ships to TestFlight and to Google Play internal testing. It is $25 a month and there is no free plan. New apps start local-first: data stays on the phone, with no accounts and no server. A backend arrives only when the app needs accounts, shared data, syncing or payments, and it brings sign-in, an API service, a Postgres database per environment and file storage. That is the data decision, made in the order this page argues for making it.
One thing no builder does for you is press submit. Uploads go to App Store Connect and TestFlight, and the review submission stays with you, which is also what guideline 4.2.6 expects. Your account, your name, your content. Code comes out as a ZIP from project settings or through two-way GitHub sync, so nothing here decides for you who owns the result.
Questions people ask before starting
An idea narrow enough to describe in one sentence, a decision about whether the data ever leaves the phone, and a paid developer account at the point where someone other than you installs it. That is the whole required list. Wireframes, a spec, a company and a logo are all optional, and every one of them is easier to get right after a first version exists.
Write the one sentence, then start
Name the person who opens the app and the one thing they do first, decide whether the data ever leaves their phone, and build the version that does only that.
Start building