For founders who built with AI
You know your industry. You know your customers. We help you know your technology.
Built your app with ChatGPT, Claude, Copilot or another AI tool? We explain its technology, data and costs in plain English, flag what could hurt you, and give your AI the exact instructions to fix it.
Your expertise built the idea. Ours makes sure it's safe to launch.
Example · Your app, explained
- Supabase
- Your database, where customer data lives. Risk Anyone with the link can read the users table.
- Vercel
- Where your website runs. Free today, about R400/month at 10,000 users.
- Stripe
- Takes payments. OK Set up correctly.
- OpenAI
- Powers the AI features. Risk No spending limit, so one busy day could cost thousands.
Vibe-coded with an AI tool? We review it.
- ChatGPT
- Claude
- GitHub Copilot
- OpenAI Codex
- OpenCode
- Cursor
- Lovable
- Bolt
- Replit
It's normal not to know
Nobody is an expert in everything
AI tools let people who understand a problem build the solution themselves. That's a good thing.
You don't need to learn databases, hosting or security. You need someone who already knows them to look at what you've built, explain it in plain English, and tell your AI what to fix. No question is too basic, and there are no lectures about how it "should" have been built.
Reviewed personally by
A software developer with 25 years' enterprise experience and 10 years of cloud architecture for local and international corporate clients.
What could hurt you
The questions that matter before real customers arrive
AI tools are good at making software that works. They don't check whether it's safe, whether it will stay up, or what it will cost when it grows.
Can strangers see your customers' data?
Logins that work but don't stop one user seeing another's data. Passwords and keys saved in the code.
What happens when it goes down?
Nobody gets an alert, backups have never been tested, and every change goes straight to live customers.
Will your bill explode?
AI and hosting costs with no limits, and a setup that gets expensive fast as you grow. We make your costs clear and point you to cost-effective options.
Questions you didn't know to ask
Instructions written for your AI
Every finding comes with a fix prompt you can paste straight into your AI tool, and a way to check it worked.
From the Production Readiness Review, you also get a rules file for your AI tool (AGENTS.md, CLAUDE.md, Cursor rules or .github/copilot-instructions.md) so it stops repeating the same mistakes. Prompts only go so far, though; automated guardrails are what make it safe.
Example fix prompt · Critical
Turn on Row Level Security for the users and orders tables in Supabase. Add policies so a signed-in user can only read and update rows where user_id matches their own ID. Show me the SQL before running it.
Example fix prompt · High
Move the OpenAI and Stripe keys out of the code into environment variables, remove them from Git history, and list the places that read them.
Illustrative examples. You run the prompts and review every change your AI makes.
How it works
Four steps, mostly done for you
Share your app
Give us read-only access. Our 5-minute guide covers GitHub, code on your computer, and app builders like Lovable, or we'll walk you through it.
We review it
We go through your app, its data, hosting and costs, the way an experienced architect would.
We talk it through
A call to explain what we found, in plain English. Weekday evenings and Saturdays are also an option.
Your AI fixes it
You get the findings, a score and instructions your AI can follow, ranked by what matters most.
Pricing
Start small. Go deeper when you need to.
Fixed prices, published up front. The Launch Check fee is credited toward a full Review booked within 30 days. Founding-client prices apply for our first 10 clients, until 31 March 2027.
| Review | What you get | Price |
|---|---|---|
| Launch Check | Your app and the paid services it relies on explained, your top 10 risks (including cost risks) and the first instructions for your AI | R3,500 |
| Production Readiness Review | The full picture: everything explained, every risk ranked, a fix prompt for each one, your costs in rands with cost-effective hosting options, and a Re-check once you've fixed them | R9,500 standard R12,500 |
| Production Hardening Roadmap | A step-by-step plan to make it solid, with checks that catch problems automatically | from R24,000 standard from R32,000 |
| Re-check | After your AI fixes it, we confirm it's actually fixed (included once with a Review or Roadmap) | R2,500 |
Also for
Businesses, agencies and investors
Running your business on an AI-built system?
Independent assurance that it's fit to rely on, before an outage, a data breach or a POPIA problem.
Production Readiness Review → AgenciesBuilding client apps with AI?
An independent sign-off before handover, so your clients know it's been checked.
Talk to us → InvestorsDon't fund a demo that won't survive production.
Technical due diligence for AI-first startups, with an executive summary for the investment committee.
Technical Due Diligence →Know your technology before your customers find its gaps.
Book a Launch Check: your app explained, your top 10 risks and your first fix instructions, for R3,500.
Your expertise built the idea. Ours makes sure it's safe to launch.
Not ready to book?
Get the free AI Application Production Readiness Checklist and run a first pass yourself.