Questions, answered.
A team of AI builders for your apps, with a dashboard you run it from. It plans, builds, tests, reviews and ships to the App Store and Google Play, then keeps going: it reads what your users report, fixes it, and ships the fix.
No. Describe the app in plain English, approve a design, and answer the occasional question. The developer accounts you need at Apple and Google are set up with you, step by step, inside Studio. If you do code, bring your existing project and the engine works inside it, with your git history intact.
iPhone and Android apps, built with Expo and React Native. A new app starts from a page of plain English; an app you already have keeps its own history. More platforms are coming, and the app's own page is where you pick them.
Your own private workspace is set up and your dashboard opens, shared with nobody. You describe your app or connect its code and Studio drafts the brief, look and feel included, for you to correct; you sign your Claude account in from the Connect page; and your builders show you their first plan before building. The full guide is on the Help page inside Studio.
As much as you want. Each app runs in one of two modes: Ask first (every piece of work is proposed to you first) or Build on its own (work runs on its own; money, public-facing changes, store submissions and open questions wait for you). Under Advanced, an optional Spend and publish without asking switch is your standing consent for money, public work and generated pictures, inside the spend caps you set; it never clears store submissions, visual sign-offs or questions from the agent. Change any of it any time.
It is the one queue in Studio where everything waiting on you collects: an approval, a question, a design to look at, a credential only you have. Answer from your phone; work picks back up the moment you do. Everything else keeps moving.
Both. The browser version is the whole dashboard, needs nothing installed, and works on any machine and any phone. The phone app is open on iPhone and Android too, through TestFlight and Google Play testing, and the install steps are on your Help page inside Studio. Either way it is the same dashboard, with a notification the moment something needs you.
Bigger changes go out as store builds: TestFlight for iPhone, Google Play testing for Android. Smaller ones are sent straight to the app as instant updates. A QA pass taps through the real app before either shipsEvery change is tested and reviewed by a second AI before either ships, and a release that needs your OK first shows up in Needs You.
Studio handles it, which is why it submits for you in the first place. It watches the review until Apple decides, then reads the notes Apple wrote and turns each reason into a piece of work. Two things come to you: the code Apple texts you, because the notes are only readable while signed in, and one tap to send the app back once the fixes are done. It never resubmits on its own.
Yes. Every plan includes five members with the access you choose. Everyone sees the same apps, progress, questions and releases at the same time, and anyone with access can comment on a release; one tap turns a comment into a piece of work for the builders.
Your own Claude subscription, your Apple and Google developer accounts, a free Expo account, and your app or your idea. Builds ship under your name, never ours. Don’t have the developer accounts? Studio walks you through creating them; Apple charges $99 a year and Google $25 once.
One app ticking over: Claude Pro, about $20 a month. Two or three apps, or one being built from nothing: Max at the 5× tier, from about $100, and the one most people want. Up to six apps: Max 20×. Seven or more: 20× plus a second Max subscription for a second member. These are our estimates from running our own fleet, not figures Anthropic publishes, so start lower if you are unsure. Running out never breaks anything: the builders pause and pick up when your limit resets.
No. The engine runs on your own Claude subscription, and we never meter tokens or resell API calls. Take the subscription over a pay-per-use API key: it is one flat price however hard your apps work, where the API bills for every word the builders read and write, typically $100 to $300 a month for a single app just being maintained. An API key still works if your company requires one.
Yes, while Studio is in beta. Every Studio bug you report that we accept takes $1 off your next invoice, and an idea you send that we ship takes $5 off, up to half your bill each calendar month. Past the cap we still want the report; it just stops taking dollars off until the month turns over. It is a thank-you, never cash, and it is for Studio itself, not for bugs in the apps Studio builds you. How the credit works.
Any time. Your apps are ordinary projects with a full history the whole time you are here. Cancel from Studio; you keep read-only access and a fourteen-day window to export everything before your workspace is removed. Within your first fourteen days, your payment is refunded in full.
Talk to the team, and a human reads it and replies. Once you are signed in, the Help page inside Studio has the full guides and the same message box.