Taking Cursor-created code to production: from the editor to an app that doesn't crash

Published on 2026-10-27 by frfrand

Checklist for deploying projects made with Cursor: architecture, Git, tests, secrets, Docker, VPS, and real maintenance.

Cursor writes code; production requires a system

The Cursor editor can help you generate a CRM, fix bugs, and document a repository. But there is no universal 'Cursor hosting' that turns any file into a public application. The way to deploy depends on whether you wrote React, Next.js, Python, Node, or a combination.

The audit worth doing before publishing

Look for .env files with keys, mocks that return false success, TODO comments in critical flows, unauthenticated endpoints, and duplicate dependencies. Run installation and build in a clean directory; what worked inside the editor does not necessarily work with different versions of Node or Python. Keep tests and Git history, review AI changes, and avoid giving the editor admin credentials if they are not necessary.

Case: CRM generated over a weekend

You created a CRM with React frontend, FastAPI backend, and PostgreSQL. Clients can be listed, but the API accepts any customer_id without checking the owning company. Before launch, you add backend authorization and tests with two different company accounts. You prepare versioned migrations and backups. Only then do you deploy frontend and API behind HTTPS and verify that the CRM resumes after restarting the VPS.

Steps that represent a completed deployment

1. Document stack, dependencies, commands, and environments. 2. Save the work in Git with revisions and automated tests. 3. Validate permissions, transactions, data, and error response. 4. Publish a staging instance with private variables and isolated database. 5. Deploy to the target server via Git and monitor logs and resources. 6. Test rollback from a previous version.

What is the difference between vibe coding and a maintainable release?

AI can generate functionality quickly; a maintainable release has an owner, traceable version, reproducible tests, and restoration procedure. You can keep using Cursor to change the code, but a change should not reach production without controls.

Frequently asked questions

### Does Cursor publish directly by itself? It depends on the tools and permissions you connect, not on the editor universally. ### Do I have to rewrite the entire project? No: first separate fixable bugs from architectural limitations. ### Do I have to manage Docker? It is not essential; it is a useful tool when it improves repeatability. ### Can I audit only with AI? It helps, but validate critical permissions and operations with independent tests.

Sources

Cursor Docs · GitHub Actions · OWASP Top 10.

Moving from AI code to ArduMaker's migration service

The goal of the future service at ardumaker.com/lovable is to study AI-generated code, adapt it to a working deployment, move services if agreed upon, and publish it on controlled infrastructure. The page does not exist yet and we are not announcing an available automatic migration tool. At ArduMaker you can consult the current VPS and service infrastructure. Each project requires diagnosis: a static landing page, an SSR with a database, and a SaaS that charges subscriptions do not need the same process.

Editorial note: October 2026. Test first on staging, document licenses, and keep verified copies before modifying production.

What the photograph shows

Network monitoring and system control. Illustrative photograph generated to accompany this guide on migrating applications created with AI; it is not a screenshot of the platform nor does it certify a specific installation.

Continue with these practical guides from ArduMaker

- Deploying an app made with Claude Code: GitHub, servers, domains, and backups - From ChatGPT or Codex to your own website: publishing AI code in production - Checklist: 30 verifications before publishing an app made with Lovable or AI

If you are evaluating your own application, compare the real backend, database, authorization, and deployment requirements before choosing the server.