Vibe code audit

Get your vibe-coded app ready for paying customers.

You built it with Lovable, Bolt, Cursor, v0 or Claude, and it works. Before real users and their data arrive, I check the code, close the security gaps, add tests and set up deployment and monitoring.

  • Fixed price and timeline in writing before you commit
  • Initial call free, by video
  • The code stays yours
Portrait of Thomas Grafenau

Thomas Grafenau

Senior full-stack developer in Austria (EU) · your contact from the first call through ongoing care

Does this sound like your app?

  • “My API keys might be sitting in the frontend code. I’ve never checked.”
  • “I can’t say for sure who can read what in my database.”
  • “Every time the AI fixes one thing, something else breaks. No test would catch it.”
  • “How the app gets onto the server is a mystery to me. If it goes down at night, nobody notices.”

These gaps are normal for an app that was built fast. AI tools build what you ask for. A launch also needs what nobody asked for: access rules, safe handling of keys, tests, a deploy pipeline and monitoring.

Six gaps between prototype and production.

The audit covers six areas. Each card shows a typical state in a fast-built prototype and what it looks like once the app is ready for production.

Logins and permissions

In the prototype
Pages and API routes check whether someone is logged in, but not whether that person may see this record or the admin area.
In production
Every route checks who is asking and what they are allowed to do. Admin functions are locked to admins.

API keys and secrets

In the prototype
Keys sit in the frontend bundle, in the repository or in its history, where anyone can find them.
In production
Keys live on the server only. Exposed keys are replaced, so old copies are worthless.

Database access rules

In the prototype
Tables are open to anyone with the public key, or the rules exist but were never tested.
In production
Each table has rules that limit users to their own data, and automated tests check those rules.

User input

In the prototype
Forms and endpoints accept whatever arrives: missing fields, oversized files, script code.
In production
The server checks every input and answers with a clear error. Uploads have limits on size and file type.

Tests

In the prototype
After each change, you click through the app by hand and hope that sign-up, login and payment still work.
In production
Automated tests cover the main flows and run on every change, before anything goes live.

Deploy and monitoring

In the prototype
Publishing is one button, and you hear about errors from your users.
In production
Every change runs through a CI pipeline into production. Errors and downtime trigger an alert.

The audit shows which of these apply to your app. You get the findings in writing.

Four steps from prototype to launch.

  1. 30 minutes · video call · free

    Initial call

    You show me the app and tell me who will use it. I ask about logins, data and payments, and tell you plainly whether I’m the right person for it.

  2. in writing, before you commit

    Fixed-price offer

    You get a written offer: what I check and fix, how long it takes, how the app will be deployed and what ongoing care would cover.

  3. in your repository

    Audit and fixes

    I go through the code and the database settings, write down what I find and fix it, highest risk first. Every change lands in your repository, where you can follow it.

  4. handover or ongoing care

    Launch

    The app goes live through the new pipeline, with monitoring switched on. Then you take over with a written handover, or I keep looking after it.

What you have at the end.

  • A written audit: what I found, what I fixed and what is still open
  • All fixes in your own repository
  • Automated tests that run on every change
  • A pipeline that takes each change from repository to production
  • Monitoring that alerts you to errors and downtime
  • A handover document: how the app is built, deployed and maintained

A fixed price to launch, a monthly fee for care.

I name the fixed price after our call, in writing and before you commit.

Getting to production

Fixed price after the initial call

The price depends on the size of the app, its user roles and the services it connects to.

After launch

Ongoing care from €500 per month

I keep the app updated, watch the monitoring and fix what breaks. Scope and monthly fee are in the offer.

Running costs for hosting, database or AI usage are listed in the offer as well.

One senior developer, from the first call to launch.

I’m Thomas Grafenau, a senior full-stack developer based in Austria, in the EU, and I’ve been building software for companies for 12 years. My daily language is TypeScript, the language most Lovable, Bolt and v0 apps are written in. My own stack is Angular, NestJS, Node.js and MongoDB; these apps usually run on React, Next.js and Supabase, and I take them to production.

I work with Claude Code every day, so I know where AI-written code tends to be thin: permissions, secrets, error handling and tests. I work on your app myself and stay your contact afterwards.

  • 12 years of software development
  • Claude Code every day
  • Austria, EU

Questions before you book.

You’re in Austria. How does that work with my time zone?

The calendar below shows open slots in your own time zone. If none of them suits you, write to office@grafenau.at and we’ll arrange a time. Most of the work needs no call: I work in your repository and send written updates.

Do we work in English?

Yes. Calls, written updates, code comments and the handover document are all in English.

What will it cost?

You get the fixed price in writing after our call, before you commit to anything, and it covers the work described in the offer. Ongoing care after launch is optional and billed monthly.

Do I have to rebuild my app?

The goal is to keep your app and make it ready for production. If a part has to be rewritten, the offer names it and says why.

Can I keep building with my AI tool afterwards?

Yes. Changes from Lovable, Cursor or any other tool run through the same tests and pipeline as mine before they reach production. The handover notes which parts need extra care when you change them.

Which apps do you take on?

Web apps built with Lovable, Bolt, Cursor, v0, Claude or similar tools, written in TypeScript or JavaScript. In the initial call we look at your stack together, and I tell you straight away whether I can take it on.

What does your app need before launch?

Pick a time for the initial call: 30 minutes by video, free of charge. You show me the app, and afterwards you know what stands between it and production. If we’re a fit, you get a written fixed-price offer.

More than 25 solutions delivered

Pick a time

30 minutes, free, no obligation

Write a message

We'll get back to you within 24 hours

Number of employees *

Your data is safe and will not be shared.

Prefer to talk directly? Call or write:+43 664 4128491office@grafenau.at