Rescue · For AI-built apps

It worked on localhost. Then you deployed.

Login loops. Supabase returning empty arrays. The AI says it's fixed and it isn't. I audit and repair apps built with AI tools — fixed price, written findings, no sales call.

Get a free Ship CheckSee what you get back

Fit

Who this is for

This is for you if:

  • You built it with Lovable, Bolt, v0, Replit, Cursor, Claude Code, or Codex
  • It works in the demo but breaks with real users
  • Auth, database rules, or deploys are failing and the AI keeps going in circles
  • You need it stable enough to charge money for

This is not for:

  • Building something new from scratch (email me instead)
  • Enterprise system integrations (see /enterprise)
  • Visual redesigns with no broken functionality

Ways to start

Three ways to work with me

Free

Ship Check

Send your repo or live URL. Within 48 hours you get five ranked findings in writing: what's broken, what's risky, what's fine. No call, no obligation.

Start here →

$

Audit

The full written report. Every issue I can find, severity-ranked, with reproduction steps and a fix estimate for each. You can hand it to any developer — or fix it yourself with your AI tools.

Request an audit →

$$$

Rescue Sprint

I fix the audit's top findings myself. Work happens in branches in your repo, delivered as pull requests you can read before merging. Includes a 14-day follow-up window on everything I touched.

Request a rescue →

The deliverable

The report is the product

Every engagement produces a written artifact: findings ranked by severity, each with the file, the cause, the reproduction steps, and the smallest safe fix. No slide decks, no vague recommendations.

Example findingIllustrative · not a client

FINDING 03 · HIGH

Row Level Security disabled on orders table.

ImpactAny authenticated user can read all orders.
FixEnable RLS + policy scoped to user_id.
Est30 min

Process

How it works

  1. 01

    Send it — repo link or live URL, plus one paragraph on what's wrong.

  2. 02

    I read it — actual code review, not a scanner.

  3. 03

    You get the findings — in writing, within 48 hours for Ship Checks.

  4. 04

    You decide — fix it yourself, hire anyone, or ask me to.

Limits

What I won't do

  • Rewrite payment or auth logic with no test coverage and no Git history — I'll tell you what has to exist first.
  • Take secret keys. Read-only collaborator access is enough for audits.
  • Pretend something is salvageable when a rebuild is cheaper — if that's the verdict, the audit says so and you've only paid for the audit.

Proof

Early access pricing

This service is new on this site. The first three rescue clients get audit-tier pricing on a full sprint in exchange for permission to publish an anonymized case study. When the case studies exist, they'll replace this paragraph.

Questions

Frequently asked

No. I work in your repo, your stack, your Git history. You should never be locked into me.
No. Read-only repo access for audits. For rescue sprints, scoped access we agree on first.
Then the audit says exactly that, with the reasoning, and a recommendation for the cheapest path forward. You only pay for the audit.
Ask. If the queue allows a rush, I'll quote it.

Intake

Send it over

One screen. Repo or live URL, a paragraph on what's wrong, and where to reach you.

Built with
What do you need?