Four ways to build it, who does the work in each, and the first step to take.
Pick the way that fits your app, the time you can give it and who will check the result. You can learn to code, build with AI, use a visual no-code tool, or get help from a developer or a partner. You can mix them. If you have not checked the idea yet, start with Is your app idea worth building?
Which way of building an app fits me?
It depends on the time you can give, how big the first task is, and whether you can tell the result works. There are four routes. Here is what each one leaves you responsible for.
| Route | Fits when | You are responsible for | Watch for |
|---|---|---|---|
| Learn to code | You have months and want to keep doing this work | Learning, debugging, running it | A slow start; building before you can judge it |
| Build with AI | The first task is small and you can say what a user does | Describing the task, checking the output, accounts and data | Results you cannot judge; the last stretch before real users |
| Visual no-code | You want visual control and accept a tool's limits | The flow, the data, the tool's rules | A tool that cannot do one thing you need |
| Hire help or a partner | Learning or the technical risk is more than you want to take | Decisions, scope, accepting the work | Unclear scope; who owns what |
Find the row where "Fits when" sounds like you, then read its "Watch for" cell before you decide. This is not a ranking. Routes can combine. You might learn enough to judge an AI result, or bring in a developer after a first version.
Founders ask this in their own words. On 17 Nov 2025 one asked whether to keep "learning how to code, or should I pay the monthly service" (post). On 1 May 2026 another was "trying to figure out the best path to actually build my app idea" (post). Each is one person's account, not research.
To choose a named tool, see Which AI app builder?. For the AI route step by step, see How to build an app with AI.
What should I decide before I pick a way?
Decide who the app is for and the one thing they should be able to do. Then answer five questions, one line each:
- The user and the useful outcome. Who is it for, and what can they do when it works?
- The data it needs to save. What must it remember?
- The device and any integration. Phone or browser, and anything it must connect to.
- Who will check it works. You, a friend, a developer.
- The time and money you can spend.
Write it on one page. That page helps on every route. If you cannot answer one of the five, that tells you what to find out before you pick a route.
For what stays in the first version and what waits, see What a first version needs. Until then, the Prototype, MVP or ready to launch? guide sets out the stages, and the App Build Planner lists features. For the cost range, see What an app costs and the App Cost Calculator.
If you have not checked the idea yet, read Is your app idea worth building? first.
What do I still have to do on each way?
Whichever way you take, you still decide what the app does and check that it works. Then each route adds its own work.
- Learn: you do the learning, the debugging and the running of it.
- AI: you describe the task, check the output and hold the accounts and data. "Without coding" is a way to start. It is not "no technical choices".
- No-code: you own the flow and the data, and you live within the tool's rules.
- Hire or partner: you own the decisions, the scope and accepting the work.
With AI, getting something running can come fast. Then the checking is yours. Does sign-in work? Is the data saved? Does it still work when a second person uses it? Test a no-code tool's limits early, so you can find out whether it supports the flow you need.
One founder wrote on 20 Feb 2026 that progress with AI slowed at deployment, integrations and bugs (post). Their "80/20" is their own phrase, not a measured rate. On a 19 Jun 2026 thread, a technical commenter warned about login, roles, billing and context (comment). Both are one person's account, not research.
Whichever route you take, know where your code, data and accounts live. Open accounts in your own name and keep the login. For the rest, see Export your code.
What is the first step on the way I choose?
Take one small step, not a plan for the whole app. These are first actions, not a schedule.
- Learn to code: build one small practice flow that has a result you can test. Stop there and see if you want more.
- Build with AI: write a short brief (who it is for, what they do, what stays out, how you will check it). Then take it to How to build an app with AI for the sequence.
- Visual no-code: test that the flow you need, and its data and access, are possible in the tool before you commit. Then pick a tool in Which AI app builder?. That guide covers named tools.
- Hire help: write the outcome, what is in scope, what is out, and how you will accept the work, before you ask anyone for a bid. The next section covers this brief.
- Partner: agree who does what, and how much time each of you gives, in writing, before you start. A guide on picking between a freelancer, a team or a co-founder is planned: Freelancer, team or technical co-founder.
Do the step, then look at the result. If it works, take the next small step on the same route. If it does not, you learn that before you spend more. You can switch routes later, with work: see Can I change the way I build later?.
When should I get help building my app?
When what you need built, or checking it, is more than you can take on. That is most likely with data, access or integrations for real users. Not every app needs a developer.
Before you talk to anyone, have four things ready:
- The outcome: what a user can do.
- What is in the first version.
- What is out.
- How you will check it works.
If you are not sure you need help, write the four things anyway. If you can fill all four, you can talk to a developer with something to check their answer against. If you cannot, that gap is what to fix first.
One founder did this with domain knowledge in hand. The Vako founder weighed Bubble's learning curve and chose help.
Once you have decided to get help, the next guide covers how to get the work done: How to get an app built. Finding and judging the people is its own guide: How to hire an app developer. Comparing a freelancer, a team and a co-founder is another: Freelancer, team or technical co-founder.
Live now: the App Build Planner writes that outline for you, and the Price and Scope Check at the end of this page works from it. For what hiring costs, see What it costs to hire someone.
What should my first version include?
One useful thing a user can do, from start to finish, plus the work that thing needs to run. Write down what waits. Leave it out of the first version, not out of the plan. A first version is not a smaller copy of the whole app. It is one complete flow: someone can start it, finish it and get something useful. Move something to the waiting list only if the promised task, the access it needs and running the app still work without it. Extra roles, screens or a second platform can wait that way. If the task needs them, they stay in.
Choose by what the user does, not by a count of screens or roles. For the full method, see What a first version needs. Until then, Prototype, MVP or ready to launch? sets out the stages.
What should I do next?
Have notes, a feature list or mock-ups? Send them for a free Price and Scope Check.
Send your brief in any form. You get one page back, usually the same working day. When demand is high it can take longer. The page shows what to cut, what it should cost as a range, how long it should take, and three questions to ask anyone who bids. It stops short of a fixed quote or a plan.
Already built something and stuck? Get a Free Build Audit →
Questions.
Do I need to learn to code to create an app?
Not on every route. Learning to code is one route. Building with AI, using a no-code tool or getting help are others. What you do need is a way to tell whether the result works, and that is the real question: who checks it. No route removes every technical choice. See which way fits.
Can I create an app for free?
You can try an idea or practise a flow with free resources. Free use for real users, hosting, store fees and a live business are different questions, and this page promises none of them. For a no-money starting point, read this guide. For dated tool costs, see What AI app builders cost.
Can I change the way I build later?
Sometimes, with work. Code, data and accounts each have their own way out, so a switch is not automatic. Keep your own accounts, and keep your code where you can reach it. Export your code covers what moving out involves.
When should I bring in a developer?
When the work, or checking it, is more than you can take on. That is most likely with data, login, integrations or anything real users depend on. Not every app needs one. Have your brief ready first: see When should I get help building my app?
Should I build for the web or for phones first?
Choose from what the user does and the device they use, not from a blanket rule. You can start on the web and add a phone app later if the task allows it. Shaman AI is a web app, with mobile work later. Mobile or web app cost covers the trade-offs.
How do beginners create an app?
Pick one small useful flow and one route. Write it on a page. Take the first step on that route. Then check the result with a real person. The first step on each route is above. For no money or no coding experience, read this guide.