Credits meter AI work. They are consumed in proportion to how much the model actually reads and writes for your request, so a small change costs a little and generating a whole application costs more. Nothing else on the platform spends them.
Free
1,000
credits every 30 days
Pro
5,000
credits every 30 days
Max
100,000
credits every 30 days
Allowances renew every 30 days from when they were granted. Monthly credits do not roll over; purchased top-ups do, and are only spent once the monthly allowance is used up. See pricing for plan details.
Credits pay for generating a whole application, not a front end. A single project can include native iOS and Android apps, a FastAPI backend, a PostgreSQL database, file storage, authentication and payments — generated together and wired to each other.
The result is committed to a GitHub repository you own, in standard React Native, Python and SQL. Host it on Appifex or move it to any provider you prefer — there is no proprietary runtime or ORM holding it in place, though generated mobile apps carry a small platform-managed folder we disclose on the comparison pages.
A credit is the unit Appifex uses to meter AI work. Credits are consumed in proportion to how much the model actually reads and writes for your request, so a small change costs less than generating an app from scratch. Every AI operation has a floor of 25 credits, and there is no per-operation cap — a larger request simply costs more.
A complete application rather than a front end. The Pro plan generates native iOS and Android apps, a FastAPI backend, a PostgreSQL database and file storage, with authentication and payments wired together — and 5,000 credits a month to build and iterate on it. The code is committed to a GitHub repository you own, in standard React Native, Python and SQL, so you can host it on Appifex or move it to any provider you prefer. Nothing is locked to a proprietary runtime.
It depends on the size of the app and how much context the model has to read, so there is no fixed price per generation. Generating a full application from a description is the most expensive operation; a small follow-up change is typically close to the 25-credit floor. Your usage is settled after each run against the real token counts, not estimated in advance.
No. When a build fails, the Fix Agent runs up to five QA and fix cycles plus a final verification pass, and that work is not billed to you. You are charged for the generation you asked for, not for the platform correcting its own output.
Deploying, previewing your app, and starting or resuming a sandbox consume no credits. Credits are spent on AI work — code generation, chat, PR review, PRD and design generation — and on managed backend hosting, which charges a floor when a backend starts or redeploys plus a smaller amount while it runs, from the same balance.
You are blocked from starting new AI work until your balance is topped up or your allowance renews. Anything already running finishes rather than being cut off part-way, so you never lose a generation that is midway through. Your apps, repositories and deployments keep running.
Monthly plan credits do not roll over — your balance is reset to the plan allowance each period. Credits you purchase as a top-up do roll over, and are spent only after the monthly allowance is used up, so buying extra credits never wastes the ones included in your plan.
Every 30 days from when the allowance was granted, not on the first of the calendar month.