Find out what's fragile or expensive in your AI-built app.

This review is for founders with a working app who are preparing to launch, take payments, or depend on it for their business. If you built it mostly by prompting and there's at least one part you'd rather not touch because you're not sure why it works, it's a good fit. It fits internal tools too: if you or someone on your team built an app with AI and the business now relies on it, the builder gets a second pair of eyes and the owner gets a clear picture of what the company is depending on. Not sure? Write to me before you book and I'll tell you whether it fits. If the discovery call shows the review isn't right for your app, I refund you in full.

My strongest ground is JavaScript and TypeScript, React, Next.js, Node.js, React Native and Python. I'm not a top expert in them, but I've worked with Angular, written Java, Ruby on Rails and C# applications, and dabbled with Go. Beyond that I consider myself language-agnostic: languages mostly differ in syntax while the semantics stay very similar, so I'm confident judging these and many more. If I feel I'm not proficient enough with your stack or your codebase, I'll tell you and refund you in full.

I'll tell you what I checked, what I found, and what I'd fix before someone depends on your app. You get a specific list in place of a general worry, including the parts that look sound. It isn't a guarantee that nothing will go wrong, and it isn't legal advice.

If my report doesn't give you something worth acting on, tell me within 7 days of our walkthrough call and I'll refund you in full. You keep the report. This covers the review; a repair sprint is a separate agreement.

The listed price covers one app with up to 1,000,000 nonblank, noncomment lines of first-party code, including relevant tests, scripts, and configuration. AI-written application code counts. Dependencies, vendored code, generated clients, lockfiles, and build output don't. If yours might be close to the limit, write to me before you book. For a larger repository or monorepo, we'll agree a self-contained portion to review or a separate quote. The report names the scope and any areas left unreviewed.

The review takes 48–72 hours from our agreed start, once I have the required access. That is separate from any wait for a start date. We agree the start date and expected delivery on the discovery call after you book.

I'm offering a limited run of early reviews at launch prices while I refine the service. Every tier includes the same scoped review, written report, and walkthrough with me. Prices advance as paid bookings fill each tier. You pay the price shown when you book.

Yes. We can agree an NDA before repository access. We'll also agree permitted tools and code handling. Read-only GitHub access and a test environment are the starting point; you don't need to make your code public or hand over production credentials.

The review identifies and prioritizes opportunities; implementing the changes is a separate step. Savings and performance gains depend on what we find and what you implement. Where useful data is available, I'll ground estimates in it and explain how to measure the result. A recommendation to keep a sound part of your app is a useful outcome too.

Yes. The report is yours to use with your coding agent or another developer. Recommendations explain the intended behavior and how to check a fix, so you have a practical starting point for implementation. You don't need to hire me for the follow-up work.

Yes, definitely. But I work through your code with it through my own lens: I zoom into the places I want to understand and judge better, and I follow my intuition wherever it leads. There's a good chance I'll read certain files myself.

I examine security concerns within our agreed scope, including access control and handling of sensitive data. The report states what I checked and where further testing is warranted. Specialist penetration testing and compliance certification are separate engagements.

I'll quote a separate repair sprint based on the findings. We agree the changes, price, timeline, and acceptance checks before I start. You receive the code changes, verification results, and a walkthrough of the decisions behind them.