Secrets ended up in the browser
API keys and service role credentials got wired into client code where any visitor can read them, usually because the chat treated environment variables as a single bucket.
Keys moved server-side and rotated
Built something with Bolt.new?
Most Bolt.new projects that land with us started as a demo that worked, grew into something a business wants to charge money for, and then ran into the parts the chat quietly skipped. We go through what it generated, keep what works, and rewrite the handful of pieces that cannot stay the way they are.
What it is
Bolt.new is an AI-driven development environment from StackBlitz, launched in 2024. A chat panel generates a React, Next.js or Astro app, installs any npm package it needs, and can ship the result to Netlify or Vercel without ever leaving the browser tab.
The trick behind it is WebContainers, which run a full Node.js toolchain inside the browser itself, so the preview you see is the real build. That makes it very good at turning a sentence into a running prototype in minutes. It is less good at being the engineer who later reads every file and asks whether it is safe to put in front of paying customers.
The honest version
No technology is all upside. This is where Bolt earns its place and where it tends to cause trouble.
Sound familiar?
The Bolt problems that bring people to us, and what you have once each one is fixed.
API keys and service role credentials got wired into client code where any visitor can read them, usually because the chat treated environment variables as a single bucket.
Keys moved server-side and rotated
Row level security was never turned on, or was turned on with no policies, so one logged-in customer can read every other customer's rows.
Policies written and tested per table
Login exists on the front door but the admin screens check nothing on the server, so pasting the URL is enough to get in.
Guards enforced on the server, not just the UI
Build command, output directory or environment variables were set in the chat preview but never carried across to the hosting provider's own settings.
Pipeline that builds green on every push
Hand-written changes vanished because the chat rewrote whole files instead of patching them, and there was no real git history to fall back on.
Moved into a real repo with proper commits
Calls that should run on a server are being made from the client, so the other service refuses them or blocks the origin.
Server routes handle the outbound calls
A single-page app shipped without the rewrite rule that sends every path back to index.html, so direct links and page reloads break.
Routing fixed at the hosting layer
The project has grown past what a prompt can usefully reason about in one go, and each small change now costs a full regeneration.
Handed off to engineers with a real editor
What we do
Call us to finish one feature, or hand us the whole prototype.
How it works
Tell us what's going wrong. We look at it, then tell you plainly what's wrong and what it will cost.
We fix the cause, not the symptom. If something is down, getting it back up comes first.
Backups, monitoring, updates and access, so the same problem can't quietly come back.
You get a plain-language write-up. Or we stay on for hosting, upkeep and new work.
Often found together
We work on the whole stack, so the problem can be wherever it actually is.
Yes. That is one of the more common things we get asked to do. We start with what you already have, keep the parts that are sound, and rebuild only the pieces that would get you in trouble as soon as a real user touches them.
Usually not without a review first. The most common issues we find are secrets shipped in the client bundle, database tables with no row level security and admin routes that trust the browser. These are all fixable, but they are worth catching before launch rather than after.
Normally it is one of a short list: the build command or publish directory never got copied across, environment variables are only set inside the chat preview, or the generated SPA has no rewrite rule for refreshes. We read the real build log and fix the actual cause.
Yes. We export the project to GitHub, clean up the history, add preview deployments on every pull request, and make sure regenerations from the chat can no longer silently overwrite a week of work.
We wire in proper authentication, add server-side session checks, and connect Stripe with real webhook handling for success, failure and refund events. The database side gets row level security that reflects who owns what.
Estimates are free. Share the project URL or a repo and you get a price before any work starts.
Free estimate
Tell us what you're running and what's going wrong. A person reads it and comes back with an estimate.
Your message is with a real person, and we'll get back to you soon.
Rather talk now? Call us: