The review your AI can't do itself.
A professional code review for apps built with Cursor, Claude Code, Lovable, and other AI tools. Security, architecture, performance — everything your AI can't evaluate about its own work.
Send your repo
Read-only access plus a brief description of what the app does.
Report in 3 days
Prioritised findings with severity ratings and specific fix recommendations.
Walkthrough call
30 minutes together. You ask questions. You leave with a plan.
What you get
Security & pen-test readiness
Authentication, input validation, secrets management, data exposure. Plus preliminary pen-test readiness checks if you're heading into enterprise security questionnaires.
Architecture assessment
Database design, API structure, will it scale past your first 1,000 users?
AI integration review
Prompt structure, eval coverage, non-determinism risk. Safe ways to add features without silently breaking the ones that already work.
Performance analysis
N+1 queries, missing indexes, asset loading — the things that slow you down at scale.
Code quality review
Error handling, edge cases, the stuff AI generates but never tests.
Report + walkthrough
Prioritised findings with severity ratings and fix recommendations, plus a 30-minute call where we go through it together so you leave knowing exactly what to do.
This is for you if…
- You built with AI and you're not sure what's under the hood
- You're about to launch and want a senior engineer to check your work
- Investors are going to ask about your tech and you want honest answers first
- You're selling B2B or into enterprise and security is a top priority
- You want the truth, not reassurance
This isn't for you if…
- You need someone to build it for you — that's what CHPTRS is for
- You want a rubber stamp for investors
- You're still at the idea stage with no code yet
What a report looks like
API keys exposed in client bundle
Your Stripe secret key and database connection string are included in the client-side JavaScript bundle. Anyone can view these in browser dev tools. Move these to server-side environment variables immediately.
No automated tests
The code works today, but there's no safety net. Change one thing, silently break three others. Start with integration tests for your critical paths: sign-up, checkout, data creation.
More findings with severity, category, and specific fix recommendations…
What AI won't tell you about your architecture
AI tools generate features, not architecture. They'll never tell you that your services are tightly coupled, your data model won't survive the next pivot, or that you're one bad deploy from a cascading failure. Each prompt gets a local answer — no one's looking at the big picture.
I've spent 12 years designing systems that survive growth, team changes, and the unexpected. This audit looks at the decisions behind the code — the ones your AI made without telling you.
Questions
What kind of codebases do you audit?
Web applications, APIs, and SaaS products. Any language, any framework. I've worked across the stack for 12+ years. If you built it with Cursor, Claude Code, Codex, Lovable, Replit, or similar AI tools, this audit is specifically designed for you.
How long does the audit take?
You'll have the written report within 3 working days of sharing access. The walkthrough call is scheduled at a time that works for both of us, usually within a week.
What do you need from me?
Read-only access to your code repository (GitHub, GitLab, etc.) and a brief description of what the app does. That's it.
What if I need help fixing the issues?
The report is designed to be actionable — you or your AI tools can fix most issues directly from the recommendations. If you want hands-on help, I offer hourly consulting at £150/hr.
Is this just an automated scan?
No. Automated tools catch syntax issues and known vulnerabilities. I find the architectural decisions, security patterns, and scaling problems that tools miss. Every finding is written by a human who's built and scaled production systems.
From prototype to production
- 01
Free self-assessment
15 common issues in AI-built apps with severity and fixes. Fix what you can yourself for free — book the audit for what's left.
- 02
£199 audit
Written report and walkthrough call. The page you're on.
- 03
£150/hr remediation
Hands-on help with the findings AI tools can't fix — architectural rework, security holes the AI itself created, anything that needs judgment beyond a single prompt.
- 04
CHPTRS for full builds
Hand the build to a senior team so you can focus on selling. For scaling what works, rebuilding what the audit flagged, or getting engineering off your plate entirely.
Your AI built it. Let's make sure it holds up.
Send me your repo link and a brief description. Report within 3 working days.