lauraq.dev · Pre-launch check for AI-built apps

Is your app ready to launch?

10 questions to answer first.

Built your app with AI? Check what could expose user data, cause unexpected bills, or stop working when customers arrive. Copy the prompt into your AI tool to review your code and find out what to check or fix first.

Useful before launch or after you’ve gone live. A starting point, not a full engineering review.

Is your app already live?Choose once and I’ll tailor the prompts below.
Step 1

Paste this into your AI tool

Open your project in an AI tool that can read its code. Start a new chat and paste the prompt. It asks for a read-only report: what’s working, what needs checking, and which changes need a careful plan.

Lovable, Bolt, ReplitUse chat or “discuss” mode, not build mode.
Cursor, Claude Code, CodexOpen the project folder, then paste into the chat.
The check-up prompt

Reviews your code without changing it. Open the link at the end of the report, or paste it into your browser, to fill in your answers here.

See the full prompt
Step 2

Check your answers

Open the link from your AI report to fill in your answers, or mark them yourself: Looks good, Not sure, or Needs fixing. Change any answer the evidence doesn’t support. Not sure is useful: it shows what still needs checking.

Open “How to check” on a question for practical checks and a prompt to investigate it.

Your answers save in this browser. Email yourself a link to continue on another device.

    Step 3

    Choose the safest next step

    Start with the first item in your results. Your next steps are grouped as Safe to try with AI, Check before changing, or Engineer review recommended. Check the evidence and work on one issue at a time.

    If your report recommends a more cautious path, follow it. If it says a change marked for engineer review here is safe to try with AI, use the verify prompt before changing anything.

    How these recommendations work

    This page uses your answers to suggest next steps and default urgency; it does not read or validate your AI report. If the report gives a different urgency and explains why, follow it after checking the evidence it cites. If the consequences or recovery are unclear, verify first.

    Fix prompt

    Asks your AI to plan one fix and stop for review if the change could affect live data, access, or payments.

    See the full prompt
    Verify prompt

    Ask for evidence before relying on a claim that something is safe.

    See the full prompt
    Step 4

    Two checks to do yourself

    A code review is only part of the picture. Try these checks before launch, or repeat them on your live app.

    1. Try it as a stranger

      A fresh account helps you see the experience without your saved settings, permissions, or familiarity with the app.

      • Sign up with a brand-new email on your phone.
      • Do the main thing your app is for, start to finish.
      • Try to break a form: leave fields empty, paste very long text, tap submit twice.
    2. Check your emails reach real people

      Check that sign-up and password-reset emails reach an address outside your team. Reading the code alone won’t confirm delivery.

      • Sign up with an address that isn't on your team, like a new Gmail.
      • The email should arrive within a minute, in the inbox, not spam.
      • If it never arrives and your app uses Supabase, it may still be on Supabase's built-in email, which only sends to your own team and only 2 emails an hour. Ask your AI how to send from your own domain.

      Careful: Sending from your own domain means adding records to it. If you already get email at that domain, like a Google Workspace inbox, have someone check the records first so you don't break it.

    Keep your progress

    I’ll email you a link back to this checklist so you can finish later. If you’ve completed all 10 questions, the email will include your results too.

    What this check can't do