ORACIS
Start a project
An engineer tests an app across a row of phones and tablets, with the code open on a laptop beside them.

Fix an AI-built app · Lovable, Bolt, Replit, Cursor, v0

Fix your Lovable, Bolt or Replit app, and get it ready for real users.

It works in the demo. Before real customers, data or payments go through it, a senior engineer reviews it, and then we finish it at a fixed price.

Photography is illustrative.

Built with
Lovable, Bolt, Replit, Cursor, v0 or Claude Code
First step
Quick Security Check in 48 hours; App Readiness Review in 5 days
You own
The code, in your repository, from the first day

What we do

It works in the demo. Real users find what the demo skipped.

AI builders are very good at the part you can see. We are a US software engineering firm; finishing the rest is the work.

  • Security and data access

    Who can read which records, enforced on the server and in the database.

  • Sign-in and accounts

    Sessions checked on the server; password reset and email that work.

  • Payments

    Stripe webhooks on the server: paying unlocks access, cancelling removes it.

  • Reliability

    Errors reported, jobs that retry, logs that explain failures.

  • Release

    Development apart from production, tests before every deploy, a way back.

  • The unfinished features

    The part where every new prompt broke something else.

How it works

A scoped review first. Then a fixed-scope completion.

Every step is priced before it starts. You decide after each one whether to go on.

Hardening a prototype for production A prototype built with an AI tool carries three open defects: secrets in the bundle, row-level security off and sessions never checked. An assessment ranks each finding with its fix, a person decides per part whether to keep or replace it, and the hardened build reaches production tested, isolated and deployed. Prototype built with an AI tool secrets in the bundle row-level security off session never checked Assessment ranked, with the fix Keep or replace decided per part Production tested, isolated, deployed secrets on the server policies tested sessions verified Hardening a prototype for production A prototype built with an AI tool carries three open defects: secrets in the bundle, row-level security off and sessions never checked. An assessment ranks each finding with its fix, a person decides per part whether to keep or replace it, and the hardened build reaches production tested, isolated and deployed. Prototype built with an AI tool secrets in the bundle row-level security off session never checked Assessment ranked, with the fix Keep or replace decided per part Production tested, isolated, deployed secrets on the server policies tested sessions verified
Fig. D-2What the tool left open, found, ranked and closed before production

  1. A free 20-minute call

    You show us the app. We tell you what we expect to find, and whether a review is worth it yet.

  2. App Readiness Review, $1,500 flat

    A senior engineer reads the code: security, sessions, data access, payments, deployment.

    You get Every finding with its severity, root cause and fix, and a fixed quote to finish.

  3. Fixed-price completion

    Security & Data Hardening from $7,500, or Production Launch from $15,000, in your repository, with tests beside each fix and every change reviewed.

  4. Launch and handover

    Separate production, a deploy you can roll back, monitoring, and an app you own outright.

Pricing

Our prices. Every step fixed before it starts.

  1. 01

    App triage call

    Free

    20 minutes

  2. 02

    Quick Security Check

    $495

    Delivered in 48 hours. The fee is credited toward the App Readiness Review.

  3. 03

    App Readiness Review

    $1,500 flat

    5 business days. The fee is credited toward any fix booked within 30 days.

  4. 04

    Security & Data Hardening

    from $7,500

    Fixed price, 1 to 2 weeks.

  5. 05

    Production Launch

    from $15,000

    Fixed price, set after the review.

  6. 06

    Ongoing care

    from $500 a month, up to $2,000 a month

    Tiered: from $500 a month for monitoring and small fixes; up to $2,000 a month for the fuller plan.

The review sizes the fixes; you get one fixed price before anything starts. How fees are credited, and the full price list.

Built with

Each AI builder breaks in its own way.

Lovable, Bolt, Replit, v0, Base44, Manus, Cursor, Claude Code, Windsurf and Firebase Studio each leave a different stack and a different set of problems. Pick yours for what usually breaks, what to try first, and how we work on it.

Proof

Who does the work, and what they have built.

The founding team has taken a product off an AI app builder before (built, not yet in paid use) and runs a production telehealth platform on Supabase. Across the systems it built and runs:

  • 20,900+ automated tests across the systems the founding team built and runs
  • 16,800+ CI runs automated builds and test runs on GitHub Actions
  • 940+ merged pull requests across the same codebases

Counted from the repositories and CI history of the systems the founding team built and runs, 22 Sep 2026; each figure rounded down. How we counted

Questions buyers ask

Before you send us the app.

Can you work with Lovable, Bolt, Replit, Cursor or v0 apps?

Yes. We work with apps from all five, and with apps built in Claude Code, Windsurf or a mix of tools. Under the tool is ordinary code, usually React or Next.js with Supabase, Postgres or Node behind it, which we write every day. The review needs the repository or an export of it.

How do I get a Lovable app to production?

Connect it to GitHub. Turn on and test row-level security for every Supabase table. Move secret keys out of the browser into server or edge functions. Use a separate Supabase project for production. Add a deploy pipeline that runs tests first, then a custom domain, backups and error monitoring. Do it yourself, or we review the app and do it with you. How we fix Lovable apps.

Is my AI-built app secure?

Nobody can say without looking, including the tool's own scanner. The same problems recur: keys in the browser, sessions never checked on the server, access decided in code the user controls, and database policies missing or too broad. That last one is how CVE-2025-48757 left 170 apps from one AI builder readable without a login in 2025. Our review checks each in your app. The five patterns we find most.

Do I have to start over?

Rarely. Most of what an AI tool produces is worth keeping: the screens, the flows, much of the data model. The review marks each part keep, change or replace, with the reason, before any work is priced.

How long does it take?

The Quick Security Check takes 48 hours and the App Readiness Review 5 business days. Security & Data Hardening usually takes 1 to 2 weeks. The rest of the completion depends on what the review finds, so the plan gives a schedule item by item. Tell us your launch date on the first call and we will say plainly whether it is realistic.

What does it cost?

The first call is free. The Quick Security Check is $495, credited toward the App Readiness Review, which is $1,500 flat and credited toward any fix booked within 30 days. Security & Data Hardening is from $7,500 and Production Launch from $15,000, each at one fixed price set after the review. Ongoing care after launch is from $500 a month, up to $2,000 a month. We never bill hourly on work nobody has sized. The full price list.

Will I own the code?

Yes. The work happens in your repository and hosting accounts from the first day. You keep the code, its full history, and the notes to run it without us.

Can I keep building with Lovable or Cursor after you fix it?

Yes, with one change: an AI tool can quietly undo a fix while building the next feature, so we set up automated checks that every change, yours or the tool's, must pass before production.

Send us the app.

The link or the repository, the tool you built it with, what worries you, and your launch date.

Get your app reviewed