Short answer: yes. Claude models write working application code today, end to end. We run a fleet of coding agents on Claude-family models that build and deploy web apps for people every day, so this answer comes from watching it happen in production rather than from a demo video.
The useful question is not "can it" but "which route fits you." There are three, and they suit different people.
Route 1: Claude.ai artifacts, for a prototype in minutes
Ask Claude in the Claude app to build something and it renders an interactive artifact: a working single-page tool, game, or mockup you can click around immediately. This is the fastest way to see your idea on a screen.
The limits are structural. An artifact is a single page running in the chat sandbox. It has no server, no database that survives, no custom domain, and no way for other people to use it at a URL you own. It is a sketch, and a genuinely useful one, but it is not a deployed app.
Route 2: Claude Code, if you are comfortable in a terminal
Claude Code is Anthropic's coding agent. It reads and edits a real repository, runs shell commands, executes tests, and builds multi-file applications of arbitrary complexity. Developers use it to build production software.
The honest catch: it assumes developer infrastructure. You need a machine with a terminal, a repo, and somewhere to deploy the result. If words like "environment variable" are new to you, this route has a real learning curve. (If you mainly want to drive it from a phone, we wrote up what runs where.)
Route 3: an app platform that runs Claude for you
Platforms like VibeKit, Lovable, Bolt, and Replit put a coding agent behind a chat box: you describe the app, the agent writes the code, and the platform hosts the result at a real URL. This is the route for "build an app for me": the infrastructure decisions are made for you.
Differences that actually matter between them:
- Which model does the building. Some platforms let you pick or bring your own Claude access; others hide the model entirely. If you have your own Anthropic key, check whether the builder accepts it, because that changes the economics.
- Where the app lives afterward. A real URL you can share, a custom domain, and an export path for the code are the difference between owning an app and renting a demo.
- What iteration costs. The first version is the cheap part everywhere. The real product emerges over dozens of follow-up requests, so per-message economics matter more than the free first build. Real numbers in how much an AI-built app costs.
"Can Claude build an app for iPhone?"
This exact question deserves a precise answer. Claude can absolutely build a mobile web app: built mobile-first, it looks and feels native, and people add it to their home screen where it launches full-screen with its own icon.
A native App Store app is a different job. It requires an Apple developer account, Xcode signing, and App Store review, which no chat-based builder does for you end to end. Claude Code can write Swift, but you own the whole distribution pipeline. If your actual goal is "my customers open it from their phones," the web app route gets you there today; the App Store route is a project.
What Claude builds well, and where projects really fail
From watching real builds across our fleet, Claude-family agents are reliably good at booking and scheduling tools, dashboards and trackers, browser games, content sites, forms, and CRUD apps of most shapes. A first working version of apps like these routinely lands in one or two messages.
Where projects actually stall is rarely code generation. The common failure points are:
- Pixel-perfect replication. "Make it look exactly like this screenshot" is the hardest request in the category, for every tool.
- Third-party services. Payments, SMS, and email need accounts and API keys only you can create. The agent can wire them, but it cannot sign up for Stripe on your behalf.
- Vague iteration. "Make it better" burns messages; "move the booking button above the gallery" lands first try.
None of these are reasons not to start. They are the difference between a five-minute demo and the several sessions a real product takes.
Try it
If you want the route-3 version of this answer, VibeKit builds and hosts the app from a chat, on the web or from the iOS app, with a free tier to see your first version live before you pay anything.