Your first app in 30 minutes, with no coding
From zero to a published app, with a link you can send by text message. The example is an order tracker for a home bakery, but the steps are the same for any small idea.
Published
Tools mentioned in this guide
- Lovable
- Builds web apps from a conversation. It has a free plan with a daily usage limit, publishes with the “Publish” button to an address ending in lovable.app, and can sync the code with GitHub.
- Bolt
- Made by StackBlitz, it builds websites and apps as you chat with the AI. It has a free plan with daily and monthly usage limits, publishes with the “Publish” button to an address ending in bolt.host, and lets you download the project or connect it to GitHub.
- Replit
- A cloud coding platform with an assistant called Agent that builds the app from your description. The free plan (Starter) has usage limits and lets you publish an app with the “Publish” button.
Details checked against each tool’s official documentation. Plans and limits change often.
Before you start, what do you need?
- A computer with a browser. You can do it on a phone, but the small screen gets in the way.
- An email address or a Google account to sign up.
- A small idea. An app that does one thing. “An app for my bakery” is too big. “A place to write down this week’s orders” is the right size.
You don’t need to install anything. Everything happens on the tool’s website.
1. Which tool should you choose?
Lovable, Bolt and Replit do the same basic thing: you talk, the AI writes the code and shows the app working next to the conversation. All three have a free plan and a button to publish.
For your first app, pick any of them. Spending hours comparing tools is an easy way to never finish any app. If you want a tiebreaker, open each site and pick the one that feels least intimidating.
2. How do you create your account?
Go to the official site of the tool you chose (the addresses are in the box above) and look for the sign-up button. All three accept email or a Google account, among other options.
The free plan has a usage limit per day or per month. For a small app, it’s enough. If the limit runs out halfway, you can wait for it to renew and pick up where you left off.
3. What should you write in your first request?
The first request has the biggest influence on the result. It needs to answer four questions:
- Who is the app for? Who will use it and in what situation.
- What does it do? The main task, in one sentence.
- What does it need? The screens, fields and buttons, as a list.
- What’s left out? What you don’t want right now. This stops the AI from inventing features.
Here’s what it looks like in the bakery example:
I want a simple app to write down orders for my home bakery. I'm the only one who will use it, on my phone.
It needs:
- a form with customer name, phone number, what they ordered, delivery date and price;
- a list of orders, with the closest delivery dates first;
- a button to mark an order as delivered.
For now it doesn't need login, payments or sending messages. It can save the data only on this device.
Use a clean look, large letters and make it easy to use on a phone. All the text in the app should be in English.
Notice the phrase “it can save the data only on this device”. It keeps the first app simple. Saving data online, so you can reach it from anywhere, involves a database and comes later.
The prompt library has other first-request templates, with a copy button.
Send the request and wait. The tool will show what it’s doing and then a preview of the app.
4. How do you test whether it worked?
The preview is the real app, working. Use it the way the baker would:
- Add two or three made-up orders, with different dates.
- Check that the list is in the right order.
- Mark one as delivered.
- Reload the page and check that the orders are still there.
Write down everything that didn’t work or looked odd. You don’t need to know why. “I clicked Save and nothing happened” is a great note.
If an error message shows up, don’t panic. See what to do when something goes wrong.
5. How do you ask for changes without breaking the app?
Ask for one change at a time and test after each one. When you ask for five things in the same message, the AI touches many parts at once and it’s hard to tell what broke what.
Be specific and say what must not change:
In the order list, show the delivery date in a highlighted style, with larger letters. Don't change anything else in the app.
The tools keep a version history. If a change made things worse, you can go back to the previous version instead of trying to patch it. It’s worth finding where that history lives before you need it.
6. How do you publish your app?
When the app does the basics right, look for the Publish button. When you publish, the tool generates a public address ending in lovable.app, bolt.host or a Replit address, depending on which one you used.
Open that address on your phone to check it. Then send it to someone you trust and ask them to use it without your help. Whatever they don’t understand on their own is your next change.
One thing to watch: in Lovable, changes you make after publishing only show up at the public address when you publish again. If the link looks out of date, that’s probably why.
Also remember that if the tool followed the request to save the data on the device, each person who opens the link has their own list, separate from everyone else’s. To share data between people, the next step is connecting a database. The Save your data online prompt helps with that.
What comes after your first app?
With an app online, you’ve already done the hardest part, which is getting out of zero. A few suggestions to keep going:
- Use the app for real for a week and write down what’s missing.
- Ask the AI to explain how the app works, using the Explain what you built prompt.
- Learn the words that come up most in the glossary, starting with prompt and deploy.
- Before you put other people’s data in the app, like customers’ phone numbers or addresses, read about what not to do with personal data.
Frequently asked questions
Can you really build an app in 30 minutes?
A small app that does one thing usually fits in that time. What takes long is wanting too much at once. If it takes more than half an hour, nothing is wrong. Finish the simple version first and improve it later.
Which of the three tools is the best?
For a first app, the difference between them matters little. All three build the app from a conversation, have a free plan and publish with a button. Pick one, finish your app, and only then compare.
Will the app work on my phone?
The app in this guide opens in your phone's browser, through a link. It does not go to the App Store or Google Play. Publishing to the stores is a different, longer process, and it isn't the goal of a first app.
Do I have to pay to publish?
No. All three tools let you publish on the free plan, with an address from the tool itself. To use an address with your own name, like yourbusiness.com, you need to register the domain and check that your plan on the tool lets you connect it to the app.
Do I need to write my requests in a special way?
No. Write the way you'd talk to a person. If something in the answer is unclear, ask the AI to explain it in simpler words.