You built a prototype with Claude or Codex.
We make it production-ready.
You shipped something that works. The next question is whether it can run in production. Secure, scalable, maintainable, and with your data intact. We make sure it can.
Code in your git. Server in your name. No lock in.
You built a prototype. Now what?
You shipped something real. So the next set of questions is real too.
Is my code reliable?
Is my code secure?
Can my users or my team get hacked?
Could my data be stolen or lost?
What happens when something breaks at 2am?
If I keep building on this, will it get better or worse?
These are the right questions to be asking. The honest answer is that most AI-built prototypes ship with problems in every one of these areas. Below is what we typically find, and what we do about it.
What changes when we are done.
Your prototype already does the thing. Everything below the first row is what we add.
Scroll the table sideways to see Production and Prototype columns.
We identify every one of these issues. We fix them. We bring your code from prototype to production.
What you get for
Equivalent agency engagements typically cost $20,000 or more. We offer the same production-grade work as a fixed flat fee, because we have systematized the process.
Production rewrite
Rebuilt on a proven stack. Typically TypeScript end to end, Postgres with Drizzle, Vite, Astro or Next.js, Tailwind, and Docker containers. The new code lives in your git from day one.
Multi pass review
Architecture, security, and QA reviewers each sign off before delivery. A senior engineer with 20 plus years of production experience approves the final code.
Security baseline
OWASP top 10 hygiene. Hashed passwords. Audit logs on sensitive actions. Server side validation of every input. No secrets in source. Rate limits and basic abuse protection.
Deployed and monitored
Your server, configured properly. Auto TLS. Nightly encrypted backups. Error tracking. Uptime checks. A CI and CD pipeline. We hand over root access and runbooks.
Third party services wired up
Resend for transactional email. Cloudflare R2 for object storage. Sentry for error reporting. UptimeRobot for uptime checks. Free tiers wherever they exist.
Runbooks
Step by step playbooks for deploy, restore, rollback, and incident response. Written so a non technical person can follow them.
Skills library
A library of AI readable instructions for your specific codebase. An AI agent can read these and continue maintaining and extending your app reliably after we leave.
What the flat fee covers
One web app, one database, the features you already prototyped. Your prototype is the spec. We rebuild exactly what it does, to production standard. As a rule of thumb, the flat $3,800 fits a prototype with:
- Around 8 data types (customers, orders, products, and so on)
- Up to 3 user roles (for example customer, staff, admin)
- Up to 15 screens
- Up to 3 third-party integrations (email, payments, storage)
What may be quoted separately
Your prototype defines how the app works, so make sure it covers the flows you need before you send it. We build it exactly as it behaves. We do not design new behaviour. These fall outside the flat fee:
- New features, screens, or flows that are not in the prototype
- Business rules or logic the prototype does not already implement
- Missing edge cases and untested flows. We flag the ones we spot, but completing them is a separate line item
- Redesigns, or scope added after the quote is accepted
Anything outside the flat fee is identified during the 24-hour review and quoted upfront, before you pay. No guessing, no surprise invoices.
Why this is worth doing once, properly.
There is a second reason to do this that does not show up in a single audit.
AI agents continue to build on whatever code base they are pointed at. If the foundation is full of shortcuts, the new code inherits the same shortcuts. Crap in, crap out. Every new feature gets a little harder to add safely, every fix is a little more likely to break something else, and the app drifts further from the standard you can show a customer.
Cleaning the foundation once changes the trajectory. From that point on, every change you or an AI agent makes starts from code that follows clear patterns, has tests, has runbooks, and has security baked in. The same AI tools that built the prototype can now keep building features without making the code worse.
Why this is worth $3,800.
Trust to ship
The code is reviewed by a senior engineer with 20 plus years of production experience, signed off after a written spec, an architecture review, a security review, and a QA review. The same review process used inside professional AI tooling shops.
AI keeps producing good code
A clean foundation means every future change, whether written by you, a developer, or an AI agent, starts from code that is easy to extend safely. The same Claude or Codex that wrote the prototype can keep adding features without making the codebase worse.
You own everything
Source code in your git. Server in your name. Third party accounts under your billing. No platform lock in. Walk away any time and everything keeps running.
How it works.
Upload
Send your project as a zip with a one sentence description.
Quote
Within 24 hours we send a confirmation, the proposed stack, the price, and a payment link.
Build
Three to seven business days. Architecture, security, and QA review passes before delivery.
Delivery
Production code in your git. Deployed application on your server. Runbooks and skills library included.
Upload
Send your project as a zip with a one sentence description.
Quote
Within 24 hours we send a confirmation, the proposed stack, the price, and a payment link.
Build
Three to seven business days. Architecture, security, and QA review passes before delivery.
Delivery
Production code in your git. Deployed application on your server. Runbooks and skills library included.
A 14 day fix window after delivery is included with every project.
What we build with.
We productionize prototypes built with
Claude Code 
Codex 
Cursor 
v0 
Lovable 
Replit 
Bolt 
ChatGPT The stack we typically rewrite on
TypeScript 
Postgres 
Drizzle ORM 
Vite 
Astro 
Next.js 
Tailwind CSS 
Docker Hosting and services we deploy with
Hetzner Default host 
Linode Alt host 
DigitalOcean Alt host 
Resend Email 
Cloudflare R2 Storage 
Sentry Errors 
UptimeRobot Uptime Yours, in full.
Nothing of what we deliver is locked behind us. You can walk away the moment we hand off, and the system keeps running.
No lock in. No vendor dependency. Walk away at any time and everything keeps running.
CRHQ. After delivery. Keep building, without becoming a developer.
We built CRHQ. It is a platform where AI agents read the skills library we installed in your codebase. The agents know exactly how your app works, how it is deployed, and how to ship the highest quality code into it.
They let a non technical person continue maintaining the application and building new features, without writing code, and without losing the production quality we delivered.
Add a CSV export to the dashboard.
Investigate why yesterday's signup flow had errors.
Draft and send the weekly customer email.
On it. Reading the skills, opening a PR, shipping when tests pass.
A senior engineer with production scars.
Andrej Šimunaj
CEO & Founder, Zero Point Studio
Every project goes through a written spec, an architecture review, a security review, and a QA review before any code reaches your git.
Common questions, before the form.
Why $3,800?
Three to seven days of senior engineer time at a fair flat rate. We charge once, instead of metering the hour, so you know the number upfront. Equivalent agency engagements typically cost $20,000 or more.
What if my app is bigger than what fits in $3,800?
The flat fee fits most prototypes: roughly 8 data types, up to 3 user roles, 15 screens, and 3 integrations. We measure the concrete structure in your zip, not lines of code or vague feature counts. If yours is over, the 24 hour review gives you the exact number that drove it and a custom quote with the same delivery guarantees. No surprises later.
What stack do you use?
Typically TypeScript end to end, Postgres with Drizzle, React on Vite, Astro for content sites, Tailwind, and Docker. Backend in Express, Fastify, or Next.js depending on the shape of the app. You see the proposed stack before you pay.
Do I need CRHQ?
No. The deliverable is fully self contained. CRHQ is an optional product if you want AI agents to keep maintaining and extending the app after we leave.
What happens to my data during the rewrite?
We work on a copy. Your original prototype keeps running until you say switch. We help you migrate the data when you are ready.
What if I do not have a git repo, server, or domain?
We set everything up in your name. You will create the accounts (we walk you through it) and you own all of it from day one.
Is my code confidential?
Yes. NDA on request, signed before we look at anything. We do not reuse client code, and we do not describe specific client projects publicly without consent.
What about long term maintenance?
A 14 day fix window after delivery is included. Beyond that, you can bring it in house, hire us to continue developing and maintaining it, or use CRHQ to keep AI agents driving the codebase.
Send us your project. We respond in 24 hours.
Zip the project folder. Send everything, even the messy parts. The only thing to skip is the node_modules folder, if you recognize the name.
Got it.
We will review your project and respond within 24 hours with a quote and a payment link. If we cannot help, we will say why, clearly, and recommend what to do instead.
Something went wrong. Please try again or email andrej@zeropointstudio.io directly.