Services AI-built apps

Your AI-built app is real software. Let’s keep it that way.

Built with Lovable, Bolt, Replit, Cursor or Claude Code and now people depend on it? We make it safe to run and keep it working, without starting over.

Getting to launch was the fast part.

AI tools are very good at producing a working app. They are less good at the parts nobody sees in a demo: who can read which records, what happens when a payment fails, where the API keys live, how to change the database without losing data.

None of that means the app is bad or that you chose wrong. It means it has reached the point where it needs the same engineering care as any other production software. That is the work we do.

What we check and fix.

Production readiness
Environments, configuration, error handling and a deploy that does not depend on one person’s laptop.
Authentication and access
Who can sign in, who can see whose data, and whether those rules are enforced on the server.
Database and migrations
A schema you can change safely, migrations under version control, and backups you have restored at least once.
Secrets and API keys
Keys out of the code and the browser, rotated if they leaked, and stored where they belong.
Tests
Coverage for the flows that matter most: sign-up, payment, the core action. Enough to change code without fear.
Security
Input validation, dependency vulnerabilities, file uploads and the other places generated code tends to trust too much.
Performance
Slow pages and queries found with real measurements, fixed where your users actually wait.
Reliability and maintainability
Monitoring, alerts and code organized well enough that the next person, or the next prompt, does not break it.

Three ways to start.

Assessment

A fixed-scope review of the code, data and hosting. You get a written list of risks ranked by what could hurt you first, and a recommendation: fix, maintain or rebuild a part.

Start with assessment

Rescue

A scoped project to fix the urgent findings: exposed keys, broken access rules, missing backups, a deploy that fails. Fixed price once the scope is clear.

Start with rescue

Maintenance

A monthly agreement to keep it updated, monitored and improving. You can keep building with your AI tools; we review, test and ship alongside you.

Start with maintenance

Not built with AI, or past the rescue stage? See ongoing software maintenance.

Questions.

Do I have to stop using Lovable, Bolt or my AI coding tool?

No. Many teams keep prototyping with the tool they know. We set up the repository, tests and review so that generated changes are checked before they reach production.

Will you tell me to rewrite everything?

Only if it is genuinely cheaper, and usually it is not. Most AI-built apps need specific fixes, not a restart. When part of the app should be rebuilt, the assessment says which part and why.

Which tools and platforms do you work with?

We work with the code and services the app runs on, whatever produced it: common setups include React and Next.js front ends with Supabase, Firebase or PostgreSQL behind them. We need access to the repository and hosting accounts. If the app cannot be exported from its builder, we will tell you what that limits.

I am not technical. Is that a problem?

No. Describe what the app does and what goes wrong in your own words. The assessment report is written for you, with the technical detail in an appendix for whoever works on it next.

Is support 24/7?

No. Monitoring runs continuously, and we respond within business-hours windows written into the agreement. We do not promise around-the-clock coverage we cannot staff.

What does it cost?

The assessment is a fixed fee agreed before any work starts, sized to your app. It also gives you a clear scope and price for any fixes or ongoing maintenance.

Stephen Way

Tell us about your app.

What it does, what it was built with and what worries you. No technical vocabulary needed. We reply within two business days.

You work with Stephen Way directly, building for the web for 20+ years. Frontend, systems, CI and testing: the work that falls between teams. Built Fjord. More about Stephen

Takes about a minute. Stephen reads every inquiry himself.

What do you need?

A few sentences is plenty: what it does and what keeps going wrong.

Is it live? Optional
When do you want to start? Optional
Budget for the first piece of work? Optional
Add more detail
Who works on it today? Optional

Free first call. Reply within two business days.