Answers / Launch safety

How do I know if my AI-built app is safe to launch?

Matt Turley, Continuum. Updated November 5, 2026.

The five checks, what you can test yourself in an afternoon, and what only a person reading the code will catch.

The short answer

How do I know if my AI-built app is safe to launch?

Check five things before real users arrive: who can read each table without logging in, whether a secret key sits in the browser code, whether one customer can reach another customer’s data, whether a payment leaves a matching order, and whether you can restore a backup. Any one failing means it is not ready. A free automated check covers the quickest of these.

The five checks, in the order that matters

AI tools build the screens that demo well and skip the dull protections underneath. These five are where launched apps most often fail, ordered by how much damage a failure does.

  • Data readable without a login. If your database answers a request from a stranger, every customer record is public. This one goes first because it ends businesses.
  • Secrets in the browser. Anything shipped in the page code is visible to every visitor, so a payment or AI-provider secret key there is as good as published.
  • Crossing accounts. Sign in as one test customer and try to open another test customer’s records by changing an address or an ID.
  • Money and records agreeing. A card charge should always leave exactly one matching order, even if the confirmation arrives twice or late.
  • Recovery. You should know where the backups are and have restored one at least once, before you need to.

What you can check yourself in an afternoon

Open the live app in a private window and try to reach pages that should need a login. Make two test accounts, put a record in each, and see whether one account can see the other’s. Place a small real payment, then check that the order exists once.

None of that needs code. It will not find everything, but it finds the failures that hurt first, and it tells you whether a deeper review is worth paying for. The free launch checklist lists the rest of the items AI tools tend to skip, each linked to the vendor documentation.

What an automated check cannot tell you

A scan looks at your app from the outside, the way a stranger would. It cannot judge whether your business rules are right, whether a rare sequence of clicks double-charges someone, or whether an admin page checks who is asking. A clean automated result is a good sign and never a clean bill of health.

Exposure is common enough to be worth testing for. In our October scan of 500 live apps built with Lovable, Bolt and base44, 177 ran on Supabase, and 31 of those 177 had at least one table that answered a request made with no login. The full method and dataset are public.

When to bring a person in

If you take payments, store customer details, or have more than a handful of users, have someone who reads code look at it before launch. A Ship Check is $299: a written, honest read of the codebase and what it needs. If the read finds problems, Ship Fix is $4,000 to $15,000, quoted fixed.

Want one concrete finding on your own app?

Paste your live app URL and your email. A short automated check runs three public-security spot-checks and sends back one specific finding and what to do about it. It is free, and it is a preview, never a clean bill of health.

Run the free is-it-safe-to-ship check. Or book a call and talk it through.

Book a call →See all services and prices →
FAQ

Straight answers

How long does a launch-readiness check take?
An automated surface check takes about a minute. Your own afternoon test with two accounts and a small payment takes a few hours. A Ship Check, where a person reads the code, is delivered in writing within two days of you granting repository access.
If an automated check comes back green, am I safe?
No. It means the quick, outside-in checks passed. It cannot see your business rules, your admin pages or rare payment sequences, so treat it as a first filter and not a verdict.
What should I fix first if several things fail?
Data anyone can read, then secrets in the browser code, then account crossing. Those can expose customers right now. Payment mismatches and backups come next because they cost you money and recovery time.
Do I need to check again after I change the app?
Yes, after any change to sign-in, payments, database rules or hosting. AI tools often rewrite those parts quietly while fixing something else, so repeat the two-account test and the payment test each time.
Keep reading
Run the free is-it-safe-to-ship checkFree. One finding on your own app.Free launch checklistThe launch items AI tools skip, each linked to the vendor docs.Ten-item pre-launch checklistRun this before you ask anyone to fix the app.The AI-built app exposure scanMethod, denominators and the full dataset.Ship CheckA human read of your codebase and an honest verdict.How to secure a vibe-coded appThe order to work in before you tell anyone it exists.Leak CheckFree automated scan of a live app, seven checks, no signup.
Sources
  1. 1. AI-built app database exposure scan, October 2026, Continuum, October 2026: Used for the figures on how many scanned Lovable, Bolt and base44 apps on Supabase had a table readable with no login. Method, denominators and the full dataset are on the page.
  2. 2. The free launch checklist for AI-built apps, Continuum: Used for the key-handling and row level security points, each of which links to the vendor documentation on that page.
  3. 3. Continuum pricing, Continuum: Used for every Continuum price on the page. The prices are read from the same list that renders /pricing.
How do I know if my AI-built app is safe to launch?