Six iOS apps. Ninety days. Already generating revenue.
That's what one developer reported after using Claude Code as their primary tool for a quarter of solo app development. At roughly one shippable app every two weeks, it's the kind of output that used to require a team.
Claude Code is Anthropic's agentic coding tool - meaning it doesn't just suggest edits, it plans, writes, tests, and debugs autonomously while you direct it - and it runs in your terminal across most major languages and codebases. It's been available since early 2025 and costs $100/month on the Max plan.
Why iOS Is a Harder Test
iOS development has historically been one of the more difficult targets for AI coding tools. Swift has less training data than JavaScript or Python, Xcode's build system adds friction that pure-code tools struggle with, and App Store submission is its own separate process. Getting six complete apps through that pipeline in 90 days - and having them clear Apple's review process and find paying users - is a more meaningful benchmark than most synthetic coding benchmarks.
The revenue part is what separates this from a hobbyist exercise. App Store approval requires functional, policy-compliant software. Actual paying users means real UX and value delivery. These weren't internal tools or prototypes.
What the Math Actually Means for Solo Developers
For independent developers, the old constraint was simple: solo output is slow output. That limited what you could build and how fast you could iterate. AI coding tools don't change the need to know what you're building - vague direction produces vague code. But they can collapse the time between "I have this idea" and "I have a working, shippable thing."
Six apps in three months is on the higher end of what people report. Domain matters a lot - some problem types are much easier for AI coding tools than others. And the developer's own experience level shapes how well they can catch AI mistakes before they compound.
But the directional shift is clear enough. A single developer with Claude Code can ship at a pace that would have required a small team two years ago, at least for apps of modest complexity. For indie developers still weighing whether to invest time in learning these tools, this kind of output data is more useful than abstract benchmark comparisons.