RLS_DISABLED
Your database is open
Supabase tables without row-level security let anyone with your public key read or change other users’ data.
Production rescue for AI-built apps
You built it with Lovable, Bolt, Cursor or Replit. We make it secure, fast and reliable enough for real users — in 7 days, at a fixed price, without rewriting it.
PreviewProduction
Click or drag
We fix apps built with
01 / The problem
AI tools are brilliant at the first version. Real users, real data and real money find everything they skipped.
RLS_DISABLED
Supabase tables without row-level security let anyone with your public key read or change other users’ data.
SECRET_IN_BUNDLE
OpenAI, Stripe or service-role keys baked into the frontend, waiting for the first person who opens DevTools.
AUTH_EDGE_CASE
Password resets, sessions, roles and redirects that hold up on the happy path and fall apart everywhere else.
WEBHOOK_UNVERIFIED
Stripe webhooks without signature checks, retries or idempotency. Paying customers without access — or the reverse.
PERF_CLIFF
Missing indexes, N+1 queries, multi-megabyte bundles and no caching. Fine in the demo, painful at launch.
PREVIEW_DRIFT
Env vars, CORS, domains, migrations and builds that only ever worked inside the editor.
02 / Free production scan
Paste your app’s URL. In about 20 seconds we map the tables, functions and pages your code reveals, and check for exposed keys, public source maps and missing security headers.
Supabase self-check
Most leaks in AI-built apps live in database rules, not in the page. Run this read-only query in your own Supabase project: in a few seconds it lists open tables, catch-all policies, leaky views and public buckets.
It reads only Postgres’s own catalog: table names and settings. It never touches your rows, changes nothing, and nothing is sent to us.
Found something red? Send us a screenshot of the result. It holds table names only, no data. We’ll reply with what to fix first and a fixed price.
Send the result03 / Process
01
Free, 20 seconds. A first look at what is exposed to the public.
Free
02
We go through your repo, database and hosting. You get a prioritised written report and a walkthrough call.
48 hours
03
We fix everything in scope as pull requests in your repo, with daily updates and plain-English explanations.
7 days
04
Monitoring, security updates and small features every month, so it stays production-grade as you grow.
Monthly
04 / What we fix
05 / Build
The same team that takes AI-built apps to production builds new ones — with security, payments and monitoring in place from the first commit.
PROTOTYPE_TO_V2
Your Lovable or Bolt prototype proved the idea. We rebuild what won’t scale and keep what works.
NEW_WEB_APP
SaaS products, dashboards, marketplaces and internal tools — designed, built and shipped.
NEW_MOBILE_APP
iOS and Android from one codebase with React Native and Expo — through to the App Store and Google Play.
Every build ships with
Fixed quote with milestones after a free 30-minute scoping call.
06 / Pricing
Scope and price are agreed before we start. The audit fee is credited in full toward your fix sprint.
Audit
$399
Delivered in 48 hours
Credited toward a fix sprint
Recommended
Fix sprint
from $1,900
7 days, fixed price
Price agreed upfront from the audit
Care
from $390/mo
For apps with real users
Cancel anytime
Our guarantee. If we don’t fix what the audit found within the agreed scope, you don’t pay for the sprint.
07 / What you get
Every finding comes with what it means for your business, how bad it is, and exactly how we’ll fix it.
Sample report · fictional app
habit-tracker.lovable.app
profiles table, including emails, with the public anon key.Add RLS · 1h08 / FAQ
No. We keep your stack and your code and fix it in place. If rebuilding a specific part is genuinely cheaper, we’ll say so in the audit — and you decide.
You do, always. Every change lands as a pull request in your repository, and we’re happy to sign an NDA before we look at anything.
Read access to your repo, limited access to your hosting and database (we’ll tell you exactly which roles), and a 30-minute call.
Yes — web apps, and iOS/Android apps with React Native and Expo. After a free 30-minute scoping call we send a fixed quote with milestones. If you already have an AI-built prototype, we usually start from it rather than from zero.
Yes. If it’s a modern web or mobile app — React, Next.js, Vite, Supabase, Firebase, Postgres, Expo — we can audit and harden it.
The audit prioritises everything. We agree the sprint scope and price before starting, so you choose what gets fixed first and nothing is billed by surprise.
We’re a small remote team working with founders across time zones. We invoice in USD or EUR.