appy.fyi finds the app; the appy Claude Code plugin builds it. That's the part I want to talk about here — the plugin, and how much of the gap between "here's a report" and "here's an app on your phone" it actually closes.
Claim it, build it
Every report on the feed ends with a "Build this" button. Click it and you've claimed that opportunity — appy.fyi hands you a build-spec.json and the exact command to run next: /appy:build <origin_play_id>. That one command is the "single prompt" — it fetches the spec, drops the project into its own subfolder, and hands off to the appy skill to do the actual work. Claim more than one, and it builds them in parallel, one subfolder each.
What "build" actually means
The skill scaffolds a real Gradle/Compose project, implements every screen and data model in the spec, wires up Google Play's In-App Review and Play Integrity APIs, writes the test plan as real tests, and generates a launcher icon from the spec's design tokens — no placeholder, no AI image call, just deterministic code. /appy:publish then walks through the one-time Play Console setup and uses the Play Developer API itself to push the build to internal testing and create the billing products.
Where it stops on purpose
It doesn't go all the way to "live on the Play Store" by itself, and that's deliberate. Trademark clearance, privacy-claim verification, the Play Console account itself, and the final promotion to production stay human-only gates. The plugin gets you to a working, tested, listable app — you decide whether to actually ship it.
That's close enough to "one prompt" that it changes what a report on the feed means: not just an idea, but a claimed, buildable starting point. Take a look at today's reports →