Step 1 · Audit
From $500
A ranked report on what it takes to be production ready, within a week. Larger apps get a fixed quote before we start.
- Six-area review
- Readiness score and ranked findings
- Harden or rebuild recommendation
- Walkthrough call
The SBD software factory
You built it fast with an AI app builder. Good. Before real users and real money go through it, a senior engineer audits it and writes a clear report on what it takes to make it production ready.
Production readiness
41/100
Harden, don't rebuild
Top findings
Built with any of these
Lovable
Bolt
Replit
Cursor
v0
Bubble
SupabaseA vibe-coded app audit is a senior engineering review of an app built with an AI tool such as Lovable, Bolt, Replit or Cursor. We check six areas against the live app and the code, and send a written report within a week, ranked by risk, that says exactly what it takes to make the app production ready. It starts at $500.
The honest part
The first two go fine. The third is where it gets expensive, because now someone else's data is inside it.
What we find in week one
None of them show up in a demo. All of them show up when strangers start using it.
Read: your Lovable app is probably publicRow-level security is off or too loose, so a table of customers, orders or messages answers anyone who asks the API directly. Nothing looks broken, which is why it ships.
The interface hides the button. The endpoint behind it still accepts the request, because the check lives in the browser and not on the server.
Payment, email and AI keys bundled into the code every visitor downloads. Someone finds them, and the bill or the inbox is theirs.
No tests and no history of why the code is shaped the way it is, so each new prompt quietly undoes an older feature.
Same app, two outcomes
60-second check
Six questions. Answer honestly. Nothing is sent anywhere.
01Can a logged-out visitor read any table through the API?
02Is any secret key (payments, email, AI) visible in the browser?
03Are admin actions checked on the server, not just hidden in the interface?
04Do you have automatic database backups you have tested restoring?
05Would you find out within an hour if the app started failing?
06Could a developer other than the original builder deploy a change?
The audit
We test the live app logged out and logged in, and read the code behind it.
Keys in the browser, open endpoints, auth checked only in the interface, and row-level security on every table.
Whether the schema survives the next thousand records: indexes, relationships, migrations and reporting.
Where it slows down at 100, 1,000 and 10,000 users, and the AI and database calls that turn growth into a large bill.
Timeouts, duplicate payment webhooks, half-finished jobs. Generated code is usually written for the happy path.
Where users get stuck, mobile layouts that overflow, and the parts that read as generated.
Secrets in the repository, one environment doing three jobs, backups, monitoring and who can deploy.
What you get
Written for a founder, not for an engineer. If the app is fine for its stage, the report says so.
After the audit
If you want us to make the fixes, they go through the SBD software factory. Every change follows the same five steps, whether it is one fix or a full rebuild.
Every flow, screen and table written down before code changes.
Small pieces, each one reviewed by a person.
Automated tests and a security check on every release.
A senior engineer approves it before it ships.
On your own accounts, with monitoring and error alerts.
Pricing
Step 1 · Audit
From $500
A ranked report on what it takes to be production ready, within a week. Larger apps get a fixed quote before we start.
Step 2 · Harden
From $1,000
We make the fixes in the report, usually in about two weeks, at a fixed price per fix.
Or · Rebuild
From $8,000
When the report says a part will not survive real users, we rebuild that part properly, on infrastructure you own.
Prices in US dollars. The statement of work fixes the price and currency before any work starts.
We build in accounts in your name. Your keys, phone numbers and data stay yours, and you can export them at any time. If you stop working with us, the system keeps running.
Security and dataWhere an automation would pay, refund or send a message on your behalf, a person on your team approves it first.
AI agents with approvalEvery automated action is logged. When one fails, it is retried and reported, and alerts go to the people you choose.
How we run systemsPlatform, AI model and phone usage bills to your own accounts at the providers’ published rates, with no markup from us.
Providers we useThe statement of work sets the scope, the price and the currency before any work begins.
How we workWork starts with the $500 audit. If your build starts within 90 days, the fee is credited in full against it.
What the audit coversQuestions
A senior engineer reviews an app built with an AI tool such as Lovable, Bolt, Replit, Cursor or v0. We check security, the data model, scale and cost, failure handling, the interface and operations, then write a report ranked by risk that says what it takes to make the app production ready.
From $500. That covers a typical MVP: one app and one database. Larger apps with several services or integrations get a fixed quote before we start, so the price never changes after you agree it.
The written report arrives within a week of the kickoff call, followed by a walkthrough call.
Read access to the code or project, a link to the live app and a test login. We never need write access to review it, and we never need your password: invite us as a team member.
No. Many AI-built apps are fine for their stage, and the report says so when that is true. A rebuild is only recommended for the part that will not survive real users.
Yes. Hardening starts at $1,000 and usually takes about two weeks. Every fix in the report has a fixed price, and you are under no obligation to take any of them.
Lovable, Bolt, Replit, Cursor, v0, Windsurf and Bubble apps, on Supabase, Firebase or a database you host. Code written by a freelancer or an earlier agency works the same way.
Your move
We'll tell you what breaks first, in plain English, and what it takes to make it production ready.
Audit, Blueprint, Build, Run. Scope and price are fixed up front, and everything sits in your name.
Start from the team that does the work, or the shape of your business.
Two-way syncs with reconciliation, retries and alerts, in accounts you own.
Systems in production, with their numbers and screens.
Field notes from building automation and AI systems, with sources.