Software Development

    Your Vibe-Coded App Is Live: 9 Things to Check Before Real Users Arrive

    Built an app with AI tools? Before real users and data arrive, check these nine things: access control, secrets, database rules, payments, backups, and more.

    8SolveAI Team5 min read

    You described your idea to an AI tool, and it built a working app. It looks great. Friends have tried it. Now you want to put it in front of real customers with real data, and maybe real money.

    This is the moment to pause. AI tools are excellent at producing something that works. They are less reliable at producing something that is safe, stable, and ready for strangers. Many of the most common problems in AI-built apps are in areas you cannot see on the screen.

    This checklist covers nine things to check before launch. It is practical guidance, not a security audit, and for anything handling sensitive data or payments you should have an experienced engineer review it. For the bigger decision about who should build what, see our article on vibe coding vs hiring a software agency.

    1. Who can see and change what? (Access control)

    The most common serious problem. Test it yourself:

    • Create two accounts. Can user A see or edit user B's data by changing a URL or an ID?
    • Can a logged-out visitor open pages or data meant for logged-in users?
    • Can a normal user reach admin pages or actions?

    If the app uses a hosted database, check that database-level access rules are turned on and correct, not only hidden buttons in the interface. A hidden button is not security.

    2. Are secrets exposed?

    Passwords, API keys, and tokens must never be visible in the browser, in public code repositories, or in client-side code.

    • Search your code and repository for keys.
    • Make sure secret keys are stored as environment variables on the server, not in the front-end.
    • Rotate any key that might have been exposed, even once.

    3. Is user input validated?

    Never trust what users type or upload:

    • Check forms, search boxes, and file uploads for unexpected input.
    • Limit file types and sizes.
    • Make sure the server validates data, not just the browser.

    AI-written code often validates in the interface only.

    4. Do logins and passwords follow good practice?

    • Use a trusted authentication service or library instead of a home-made login.
    • Require strong passwords and offer password reset by email.
    • Consider multi-factor authentication for admin accounts.
    • Make sure sessions expire and logging out works.

    5. Are payments handled safely?

    • Use a reputable payment provider and its hosted checkout. Never store card numbers yourself.
    • Verify payments on the server, using the provider's confirmed notifications (webhooks), not just a "success" page.
    • Test refunds, failed payments, and duplicate attempts.
    • Make sure prices and totals are calculated on the server, so users cannot change them.

    6. Are there backups and a way to recover?

    • Turn on automatic database backups, and test restoring one.
    • Know where your code lives and keep it in version control.
    • Know how to roll back a bad update.

    "It works on my computer" is not a recovery plan.

    7. Will you know when something breaks?

    • Add error monitoring so you are alerted when users hit problems.
    • Set up uptime monitoring for your site.
    • Keep logs that help you diagnose issues, without storing sensitive data in them.

    Without monitoring, you learn about outages from angry customers.

    8. Is it fast and stable under real use?

    • Test on a real phone and a slow connection.
    • Try with more data than your test data. Do lists and searches slow down?
    • Check what happens when many users act at once, or when an external service is down.
    • Look at your hosting limits and costs as traffic grows.
    • Publish a privacy policy and terms that match what the app actually does. Take advice from a qualified professional for your market.
    • Be clear about what personal data you collect and why, and offer a way to request deletion.
    • Confirm that you own the code, the database, the domain, and the hosting accounts.
    • Check the licences of any templates or libraries used.

    A quick pre-launch routine

    1. Test with two accounts and try to break access rules
    2. Search for exposed keys
    3. Test bad input and file uploads
    4. Verify login and password reset
    5. Run a payment test end to end
    6. Take and restore a backup
    7. Turn on error and uptime monitoring
    8. Test on a phone and under load
    9. Publish privacy and terms pages

    When to get professional help

    Bring in an experienced engineer if your app:

    • Handles payments, health, financial, or other sensitive data
    • Has many users or a lot of traffic
    • Has complex rules or integrations
    • Will be the foundation of a business

    A short code and security review is far cheaper than a breach, a data leak, or an emergency rebuild.

    Do not throw the app away

    An AI-built prototype is a real asset: it has already proved the idea and clarified the requirements. A good partner can review it, fix the risky parts, and rebuild only what must be rebuilt. See our guide to what makes a good MVP.

    We can review and harden your app

    We review AI-built apps for security, stability, and scalability, and help you launch safely. See our startup product development service and software development services, or book a free consultation.

    Have a project in mind?

    Tell us what you want to build and we will suggest the simplest path to get there.

    Keep reading

    All articles