Field notes on
fixing AI code
Expert insights on fixing AI-generated code and shipping production-ready software.
How to Prompt for Less Buggy Code
Better AI code prompts stop bugs before they start. Practical prompt engineering coding habits for Cursor, Bolt, and Lovable users.
All articles
The Hidden API Keys in Your AI-Built App (And How to Find Them)
Exposed API keys are the most common serious flaw in AI-built apps. Here's where they hide and how to find yours today, even if you can't read code.
Rebuild or Rescue? When an AI App Isn't Worth Saving
Rebuild vs refactor for AI-generated apps: an honest framework for when to rescue your vibe-coded app and when a rewrite actually saves money.
AI App Accessibility: The Risk Nobody Mentions
AI app accessibility gaps are a growing legal exposure. Here's why AI-built UIs fail WCAG and ADA standards, and how to fix it before launch.
How to Hand Off a Vibe-Coded App to a Real Dev Team
Ready to hand off AI code to a dev team? Here's what founders need to prepare, what a good team asks for, and the red flags to watch for.
The Vibe Coder's Guide to Not Losing Your Data
App data loss is more common than founders think. Here's how AI-built apps lose data, why 'it's in the cloud' isn't a backup, and what to check today.
Can You Trust AI to Handle Payments? Stripe Red Flags
AI-generated Stripe integrations often skip webhook checks, trust client-side prices, and leak secrets. Signs your AI payment integration isn't safe.
Supabase and AI Tools: The Database Mistakes That Bite Later
Supabase is the default backend behind AI-generated apps, and Supabase RLS mistakes are the reason so many of them leak data. Here's what to check.
How to Add Authentication to a Vibe-Coded App the Right Way
Adding auth to an AI-built app? Here's what makes AI app login security different from login that just works, and where AI tools get it wrong.
Why Your Lovable App Is Slow (And How to Fix It)
Lovable app slow? Here's why AI-built apps lag under real traffic and how to fix ai app performance without a full rebuild.
Replit Agent Code: The Problems We See Most
Replit Agent bugs we see most often in fixed apps: leaked secrets, over-eager scaffolding, unsafe database defaults, and deployment surprises.
How Much Does It Cost to Fix a Vibe-Coded App?
What's the real cost to fix an AI-generated app? We break down what drives the price, typical cost bands, and the red flags that make a job expensive.
The Future of Vibe Coding: Will AI Tools Fix Their Own Bugs?
Will AI coding tools learn to fix their own bugs? An honest look at where vibe coding is heading in 2026, what's improving fast, and what still needs a human.
AI Code Review Tools vs. Human Code Review: What Actually Works
Automated code review tools like SonarQube, Snyk, and CodeClimate catch real problems, but they miss the ones that hurt most. Here's what each is good for.
The Vibe Coder's Checklist: 20 Things to Check Before You Launch
A practical pre-launch checklist for AI-built apps: 20 security, performance, reliability, and compliance checks you can run before real users arrive.
Vibe Coding and GDPR: What Non-Technical Founders Need to Know
AI coding tools don't think about data protection. Here's what GDPR expects from your app, the violations we see most in AI-generated code, and a plain-English checklist.
5 Signs Your Vibe-Coded App Needs Professional Help
Not sure whether your AI-built app needs a real developer? Here are five honest signs that DIY fixes have run out, and what to do next.
Is Your AI-Built App Secure? A Non-Technical Guide to Vibe Code Security
AI generated code security risks explained in plain English, from exposed API keys to missing auth, plus the vibe coding security checks any founder can run today.
What Is Vibe Coding — And Why Your App Probably Has Bugs
What is vibe coding, and why does it lead to ai coding bugs? A plain-English explainer on the vibe coding problems that show up once real users arrive.
Why AI Coding Tools Keep Making the Same Mistakes
AI coding tools are useful, but they repeat the same bugs because they lack runtime feedback, full context, and product judgement.
From Vibe Code to Production Code: A Step-by-Step Upgrade Guide
A practical roadmap to improve AI-generated code, harden a Cursor-built app, and make vibe code production ready before launch.
The Real Cost of Free AI-Generated Code
AI-generated code feels free until bugs, rewrites, security fixes, and lost users start costing real money. Here's how to estimate the risk.
How to Review AI-Generated Code (Even If You're Not a Developer)
A practical AI-generated code review checklist for non-technical founders using Cursor, Bolt, Lovable, v0, or similar tools.
The Non-Technical Founder's Guide to Code Quality
Five signals of code quality you can check yourself with no coding knowledge required, plus free tools and when to call a professional.
How We Fixed a $50K App Built Entirely with Cursor
A case study: a seed-funded SaaS built with AI coding tools, breaking under real users. What we found, what we fixed, and the results.
Cursor vs Bolt vs Lovable: When to Fix vs When to Rebuild
A bug-fixer's perspective on code quality from the most popular AI coding tools, and the framework for deciding whether to fix or rebuild.
The 10 Most Common Bugs in AI-Generated Code
A tactical guide to the bugs we find most often in Cursor, Bolt, Lovable, and v0 codebases, with code examples showing the fix.
Why Your Vibe-Coded App Breaks in Production
Your AI-built app works perfectly in development. Here's why it falls apart in production, and what to do about it.
Need help with your AI-generated code?
We fix bugs in vibe-coded apps and make them production-ready. Free code assessment — no commitment required.