Draft, not yet published
How much does it actually cost to make a vibe-coded app production-ready?
7 July 2026· 6 min read · by Stackbastion
Your app works. Now you want real users on it, and everyone’s telling you it needs to be “production-ready” first. Nobody will tell you what that actually costs. The quotes you get range from “a weekend” to “€20,000,” which is useless. Here’s an honest breakdown of what the work is and what it really costs, in time and money.
What “production-ready” actually means
Before pricing anything, get specific. Production-ready isn’t one thing. It’s a list, and you might not need all of it on day one.
- Backups you can restore. A tested backup with point-in-time recovery, not just platform snapshots.
- Secrets handled properly. No API keys in the code, everything in environment variables, nothing leaked.
- A real database. Often moving off a bundled platform database onto managed Postgres you control.
- Error handling and monitoring. You find out something broke before your customers do.
- Basic security. Row-level access rules so user A can’t read user B’s data, rate limiting, HTTPS.
- A safe way to deploy changes. So one bad update doesn’t take the whole thing down with no rollback.
- Legal basics. If you’re in the EU or serving EU users, a data processing agreement and somewhere compliant to store the data.
You don’t need every item to launch to 10 users. You do need most of them before you take money or hold customer data.
The cost drivers nobody mentions
The price swings on a few things. Know which apply to you.
- How much data you already have. An app with no users is easy to fix. An app with 5,000 customers and live orders means every change has to be careful and reversible, which takes longer.
- How the AI tool built it. Some generated code is clean enough to harden in place. Some is tangled enough that parts need rebuilding. You can’t know which until someone reads it.
- Whether you’re moving off the platform. Staying on Lovable and hardening is cheaper than migrating to your own hosting. Migration adds real hours.
- Compliance scope. EU customer data raises the bar: DPA, data residency, breach process. Not hard, but not free either.
Realistic cost ranges
Here are honest ranges for the three common situations. Treat these as ballparks, not quotes.
DIY, you do the work
Money: €20 to €150 a month in services. Time: 20 to 60 hours to learn and set up.
If you’re technical enough to follow guides, you can do most of this yourself. Managed Postgres, a backup tool, a monitoring service, and a host add up to maybe €20 to €150 a month depending on size. The real cost is your time: expect a few full days to learn backups, secrets, and deployment if it’s new to you, plus ongoing hours when something breaks. This is the cheapest in cash and the most expensive in attention.
Hire a freelancer for a one-time hardening
Money: €800 to €5,000 one-time, plus monthly services. Time: 1 to 4 weeks calendar.
A good developer does a discovery, then hardens the app: backups, secrets, database, monitoring, deploy safety. The range is wide because it depends entirely on how the app was built and how much data it holds. A clean, small app might be a few days’ work. A tangled one with live customers can run into weeks. Add the monthly service costs on top, and remember you now own the ongoing maintenance and the “who fixes it at 2am” question.
Managed hosting and ops
Money: roughly €39 to €249 a month, depending on tier. Time: near zero of yours.
This is the “someone else runs it” option, and it’s where we sit, so read this as one data point rather than the only answer. Stackbastion’s pricing runs from €39 a month on Launch to €249 on Business, with tested backups, PITR, monitoring, and a named human on call included. The trade-off is honest: you pay every month instead of once, but you don’t carry the ops burden, the on-call risk, or the “my one freelancer went on holiday” problem. For a badly broken app that needs urgent rescue, there’s a fixed-price Rescue engagement from €490 rather than an open-ended hourly bill.
The cost of doing nothing
Skipping this has a price too, it’s just deferred and lumpy.
- A wiped database with no backup can end a small business. There’s no range on that, it’s the whole thing.
- A leaked API key can run up thousands in charges on your cloud or AI bill before you notice. We’ve seen a single exposed key cost real money.
- A GDPR complaint with no DPA and no compliant hosting turns into legal cost and, in a bad case, a fine.
The point isn’t fear. It’s that “production-ready” spending is buying down these risks, and the cheapest option today can be the most expensive one over a year.
Or, we do it for you
Not sure which situation you’re in or what it’d cost? A free Stackbastion audit tells you the real state of your app and what hardening it would actually take, with no obligation. You can take that report to a freelancer, do it yourself, or hand it to us.
FAQ
Why can’t anyone give me a fixed price up front?
Because the cost depends on how the app was built and how much data it holds, and nobody knows that until they read the code. A fixed quote before discovery is a guess dressed up as a number, and it usually goes one of two ways: they pad it to be safe, or it runs over and you argue about it. A short paid discovery first gives everyone a real number.
Is monthly managed hosting worth it over a one-time fix?
It depends on how much the app matters and whether you want to own the ops. A one-time fix is cheaper in total cash if nothing ever goes wrong and you’re happy handling incidents yourself. Managed hosting costs more over a year but removes the on-call burden and the single-person risk. If the app is a hobby, do it once. If a business depends on it, the monthly cost buys you sleep.
What’s the minimum I should spend before taking real users?
Backups you’ve tested, secrets out of the code, and basic access controls so users can’t see each other’s data. That’s the floor, and you can hit it cheaply, even for free on your own time in a weekend if you’re technical. Everything above that (monitoring, migration, compliance) is worth doing but can wait a little. The floor cannot.
Does moving off the AI platform cost extra?
Yes, migration adds hours because the data and code have to move and be re-tested somewhere new. Staying on your current platform and hardening in place is cheaper. Migrate when the platform is genuinely holding you back (connection limits, cost, no control over backups), not just because it feels more “real.” Match the spend to the actual problem.