- Finishing
- AI-generated code
- Launch
How to finish an AI-built app that’s stuck at 80%.
The last 20% of an AI-built app is integrations, edge cases and deployment. Take an honest inventory, wire real payments and auth, and get it live.
Read
9 min read
Fix my v0 app
v0 generates good-looking screens — and stops there. What it hands you is usually screens filled with placeholder data, no real connection to a database, and some code running in the wrong place (Next.js splits code between the server and the visitor’s browser, and v0 doesn’t always get that split right). The fixes are specific: connect real data, put the code where it belongs, and make the Vercel publish succeed for the right reasons.
Where v0 projects usually break
- 01
Works in preview, fails to publish on Vercel
Small mistakes the preview tolerated but the real publish doesn’t — often one wrong reference or a missing setting.
- 02
Code running in the wrong place
Next.js runs some code on the server and some in the browser. v0 often mixes them up, and the usual quick fix — adding ‘use client’ everywhere until the errors stop — quietly breaks how data loads.
- 03
Placeholder data that never became real
Lists typed straight into the screen so the design looked right. There is no database behind them, and no ‘loading’ or ‘something went wrong’ state.
- 04
Forms that go nowhere
A beautifully styled form with no server action behind it, or one that succeeds visually and saves nothing.
- 05
Login and payments as pictures only
Login screens and pricing pages that display, with no real accounts, no protected pages and no payment provider connected.
Try these first
- Read the Vercel build log to the first error — not the summary, the first red line.
- If you have a developer handy: run `next build` on their machine. If it fails the same way, the fix is in the code, not in Vercel.
- Search the code for ‘use client’. If it sits at the top of almost every file, the server/browser split has been papered over — a reliable sign the data loading needs proper wiring.
- Look at any list on a screen and ask: where does this come from? If the answer is ‘it’s typed into the code’, that is a data source still to be built.
- In Vercel, open Settings → Environment Variables: every setting the app needs must exist there, under the same name.
How I work on v0 projects
The screens v0 produced are usually the part worth keeping, so I keep them. I add what is missing underneath: a real connection to a database, logins that actually protect pages, payments wired to a real provider, and the loading and error states the placeholder data never needed.
Code gets put where Next.js expects it, rather than wherever the errors went quiet. You get a repository that publishes cleanly on Vercel and a note on what each fix was for.
Common questions
Can you work with Cursor, Lovable, Bolt, Replit or v0 projects?
Yes — all of them, plus Claude, Copilot and ChatGPT output. I don’t much care which tool generated the code. What matters is the application underneath it: the architecture, APIs, database, dependencies, and where things are breaking. Lovable and Bolt projects tend to need deployment and backend work; Cursor and Copilot projects more often need the seams between generated pieces checked.
Can you build something from scratch?
Yes — mobile, web and AI-powered products. It's a different engagement from a rescue: we start with what the thing actually needs to do, and I'll usually push to get a small version in front of real users before building the rest. If your idea is better served by something simpler than what you're picturing, I'll say so.
Can you help with deployment?
Yes, and it’s one of the most common reasons people get in touch. "It works locally" is almost always an environment, configuration or build problem rather than a code problem, and those are findable.
How much does it cost?
It depends on the work, so I don’t publish a price list. After a first look I quote a fixed price for the diagnosis, and any fix is scoped from the findings — so you know the cost before any work begins, and nothing changes partway through.
The service
Finish My AI-Built App
Most of it exists. The last stretch is the part where the AI stopped being useful — the integrations, the edge cases, the settings screen nobody wants to build, the deployment that never quite worked.
Related problems
- Finish my AI-built appMost of the app exists. The last stretch is integrations, edge cases and deployment — the parts the AI stopped being useful for. Here's how it gets done.
- Fix my AI-generated appYour AI-built app is broken and the prompts aren't helping. Here's how the problem gets diagnosed and fixed without rewriting what already works.
- Fix my Lovable appLovable app broken, stuck in a fix-one-break-another loop, or failing after publish? How Lovable projects get diagnosed and fixed without rebuilding.
Further reading
- Deployment
- Production
It works locally but not in production. Here’s the checklist.
Nine differences between your machine and production, in the order they break: env vars, build output, ports, database, CORS, paths, time, memory and secrets.
Read
9 min read
Your AI-built app doesn’t need another prompt.
It might need an engineer.
If you’ve spent hours asking AI to fix the same problem and you’re going in circles, stop fighting the code. Send me what you’ve got — I’ll figure out what’s wrong and tell you what I’d do next.
No cleanup. No judgment. The messy version is usually the useful version.