How to make an app without coding (by describing it in chat)
Most “make an app without coding” guides hand you a drag-and-drop editor and three tutorials. You still end up doing the developer’s job — choosing databases, wiring logins, connecting “workflows” — just with more clicking.
There’s a simpler honest answer now: describe the app, and let AI build it. Here’s what that actually looks like, step by step, including the parts that are genuinely free and the limits nobody puts in the headline.
What you need
- The idea, in one or two plain sentences — “an app where my customers book a slot”, “a study tracker for my students”, “a checklist my family shares”.
- A phone or a laptop with a browser. That’s the whole toolkit.
You do not need: an AI subscription or API keys, a database, a developer account, or an app-store review. That’s four gates removed — and they’re precisely the four where most app ideas die.
The five steps
1. Say what you want. On whacka, you start by typing (or dictating) your idea in a chat. Don’t write a spec — write it like you’d explain to a capable friend. Details you forget now can be added later; nothing is locked in.
2. Watch the first version appear. The AI writes a real app — screens, buttons, data, logic — and you see it in a live preview next to the chat, usually within minutes. This first version will be roughly right and wrong in specific ways. That’s normal and useful: now you have something concrete to react to.
3. Refine in plain language. “Make the list show newest first.” “Add a photo to each item.” “Only I should see the totals.” Each message updates the app in front of you. This loop — look, say, look again — is the actual “development process”, and it’s the part where your taste and your knowledge of your own situation do the work no template can.
4. Publish. One tap gives your app a permanent link of its own on the web. Anyone you share it with can open it instantly in their browser — no downloads, no accounts on our side, nothing to install.
5. Install it like an app (optional). Your published app can be added to a phone’s home screen, with its own icon and full-screen feel. For you and your users it behaves like an app from a store — it just skipped the store.
What can you realistically build?
Real apps with real moving parts, not just pages: things that save data (orders, bookings, records, scores), know who’s who (your users can sign up and log in to your app), update live (lists and chats that refresh for everyone), send notifications, and even use AI inside (your app can summarize, translate, generate — without you touching any AI setup).
Concretely, the patterns we see built most: booking and appointment pages, order links for small sellers, trackers for students or clients, family checklists and habit apps, community boards, little tools people got tired of doing in spreadsheets.
What it costs
Starting is free, and the free tier is enough to build, iterate, and publish a real app. Building consumes daily free credits (they refresh); if you build a lot, paid plans add more. You are never charged for “seats”. One thing to know: on the free plan a published app stays online for 14 days after each publish (publish again and the clock restarts); paid plans keep it online for as long as you’re subscribed.
The honest limits
- It starts as a web app. It installs to the home screen from its link and feels native — for most personal and small-business apps that’s a feature, because your users skip the download wall entirely. If you do want a Google Play / App Store listing, that’s a separate step on the Pro plan.
- It’s not for everything. A stock-trading platform, a game studio’s next release, deep hardware integrations — wrong tool. One-screen-to-a-dozen-screens apps that organize real life and small business — right tool.
- The first version won’t be perfect. The product is the refinement loop. People who get great results treat the chat like directing a very fast assistant, not like a vending machine.
The gap between “I’ve been meaning to build this” and “here’s the link” is now measured in an afternoon. The only step that can’t be done for you is deciding to describe the thing.