Build Apps With AI
You don't need to be a software developer to build an app anymore.
People are using tools like Claude, Cursor, Replit, Lovable and Codex to turn ideas into working software, sometimes in days or weeks. But getting from an idea to something people can actually use still takes judgment, testing and a willingness to figure things out as you go.
On App Minded, I talk to people who have done it.
Some had worked around technology before. Some started with no coding background at all. They have built iPhone apps, web apps, AI tools and products around problems they found in their own lives.
These are their stories and what they've learned along the way.
Can you really build an app with AI?
Yes.
That doesn't mean typing one prompt and watching a finished business appear.
What has changed is who can get started.
Jocelyn Greenlee is an SEO strategist with no coding background. She had been given three chickens and couldn't find a suitable coop, so she started researching how to build one. The information she found was complicated enough that another idea occurred to her: what if she built an app that could look at a photo and create the build plan?
She went home and started building Buildara with Replit.
She estimates that after two focused days of roughly eight hours each, she had something she considered viable enough to ship.
It wasn't flawless. Buttons broke. Changes caused problems elsewhere. She sometimes had to research technical concepts just to explain what she wanted.
But she had a working product.
And Buildara eventually made it into the iOS App Store.
From everyday problem to App Store
A pattern comes up repeatedly on App Minded: the idea starts with something annoyingly ordinary.
Kayla Orozco is a mother of three. Every school year she received three different school-supply lists and had to combine them herself.
She was tired of doing it.
Kayla worked with AI professionally as a conversation designer, but she wasn't a developer and had never built her own app.
She started working through the idea with Claude and decided to make it an iPhone app.
Eight and a half weeks later, School Supplied was in the App Store.
Christina Moore's idea came from book clubs.
She had a finance background and described herself as knowing nothing about coding, web design, graphic design or UX when she started. She entered a 48-hour Lovable hackathon with an idea for solving the logistical mess around running a book club: RSVPs, locations, questions, contributions and the endless group chat.
That became Swan Society.
Deepti's starting point was much more personal. After kidney failure, eight months of dialysis and a kidney transplant, she kept running into the same problem: what could she actually eat?
There was plenty of information available. That was part of the problem. She could spend three or four hours looking through recipes without making a decision.
During a hackathon, she built Nourishly around that problem. Her team won first prize.
Different apps. Different backgrounds. Same starting point.
There was a problem close enough to the builder that they understood why it mattered.
What tools are people using to build apps with AI?
There isn't one stack everyone on App Minded uses.
Guests have built with:
- Claude
- Claude Code
- Cursor
- Codex
- Replit
- Lovable
- Bolt
- Base44
- Vercel
- AI-assisted automation and API tools
Some stay in browser-based builders. Others gradually move into more technical tools as their apps become more complex.
Barbora, a biochemist by training, moved from business automation into building Xroad Studio, a content creation and distribution product. She has tried Cursor, Codex and several browser-based AI builders, but prefers coding agents such as Claude Code when she wants more control over a production product.
Chiara Barden uses Claude alongside Replit. Instead of asking one AI to do everything, she created a virtual software team with different roles covering areas such as security, frontend development, UX and market research. She works through the problem with that team in Claude, then takes more developed instructions into Replit.
Cynthia Vasquez built an early version of one of her apps with Lovable, rebuilt it in Replit and eventually self-hosted it through Vercel.
The tool is only part of the process.
Do you need to know how to code?
Not necessarily.
But building an app will make you learn things.
Christina Moore didn't come out of her first serious build knowing a programming language. What changed was her understanding of how software fits together.
That's an important distinction.
AI can translate a lot of what you want into code. You still have to understand your product well enough to know when the result is wrong.
You start learning what a database does because you need one.
You learn about APIs because your app needs to talk to something else.
You encounter authentication, security, deployment, App Store requirements, payments and bugs because your product eventually forces you to.
Several App Minded guests describe themselves as non-technical and then, twenty minutes later, are discussing APIs, MCPs, databases and deployment.
They learned by building.
AI makes building faster. It doesn't remove product decisions.
One of the more useful stories from the podcast is about an app that didn't work.
El Wong built an app for organizing pickleball games. It handled the court, time, payments and reminders.
Technically, the idea worked.
Then she got feedback.
People didn't particularly want another app. They were already organizing their games through text messages. The existing solution was messy, but it was familiar and good enough.
That's a product problem, not a coding problem.
AI can dramatically reduce the cost of making something. It doesn't guarantee that anyone wants it.
The easier building becomes, the more important this question gets:
Should this exist?
Start smaller than you want to
Building with AI makes it dangerously easy to add things.
Another screen. Another feature. Another clever use of AI.
That doesn't mean you should.
The first version needs to prove the central idea.
Can the person you're building for understand it?
Does it solve the problem?
Will they use it again?
Barbora's experience building a product and working with paying customers has made her selective about feedback. Lots of people will suggest features. Her response is essentially: would you actually use the product if we built that?
Feedback from someone paying and using the product regularly carries a different weight from an interesting suggestion.
Getting something built is no longer the whole challenge
This comes up constantly on App Minded.
Distribution is hard.
Barbora had beta testers lined up before Xroad Studio launched. Her first customers came through direct outreach and existing communities. One of the earliest came from a WhatsApp group where someone happened to mention needing exactly the type of product she was building.
El describes building as one half of the job and marketing as the other. She has built multiple products but still finds distribution harder than building.
Cynthia has the same problem from another direction. She can build quickly enough that she now has to stop herself from immediately turning every idea into an app. Her focus has shifted toward testing and promoting what she has already made.
This is one of the reasons App Minded isn't just about coding.
An app nobody discovers is still an app nobody uses.
What does building an app with AI actually look like?
Across the people I've interviewed, the process tends to look something like this:
- You notice a problem.
- You work out whether an app is actually a sensible solution.
- You make the first version much smaller than the version in your head.
- You use AI to plan and build it.
- You test constantly.
- Things break.
- You learn enough about the technical pieces to diagnose what is going wrong.
- You put it in front of people.
- You find out which of your assumptions were wrong.
- You change it.
Then you have to figure out how anyone is going to find it.
That's much less magical than the AI demos.
It's also much more interesting.
Listen to people who are actually building
App Minded is a podcast about what happens between having an app idea and getting that app into people's hands.
The guests include first-time builders, founders, marketers, developers and people who never expected to build software at all.
We talk about the tools, but also about the parts that don't fit neatly into an AI demo: bad first versions, bugs, App Store submissions, validation, pricing, customers, marketing, distribution and what people would do differently next time.
If you're wondering whether you can build an app with AI, start with the stories of people who already have.