SEO for vibe coded businesses
You built it yourself with Lovable, Cursor, v0, or Bolt, and it looks better than your competitors’ agency sites. Then you searched for your own business and nothing came up. ChatGPT has never heard of you either. I fix exactly that.
A new kind of business owner
A whole wave of businesses now ship their own websites by describing them to an AI. It is a real, new industry, and it comes with one problem nobody warned you about: the tools were built to make a site look right, not to make it findable. The story is the same almost every time.
A weekend, a prompt, a few revisions. No agency, no developer, no six-week timeline.
Clean layout, real copy, working forms. Better than what most competitors paid thousands for.
Nothing. Or the homepage alone, with every other page missing. ChatGPT does not know you exist.
The builder says the site is live. The SEO tools say it is fine. Both are looking at the wrong thing.
I have watched a full marketing site get indexed as a single URL because of one tag the builder wrote by default. The fix took an afternoon. Finding it took someone who reads what the crawler reads. That is the whole job, and it is the only job I do.
The three failures
These are not exotic problems. They are the default output of tools that were built to make a site look right, not to make it findable. Every one of them is fixable without touching your design.
The builder emits a single canonical tag pointing at the homepage, on every route. Each page on your site is telling Google, in writing, that it is a duplicate of the homepage. Google obeys the tag, files your pages as copies, and indexes the homepage alone. Rebuilding does not fix it, because search engines are obeying the instruction, not the code.
<link rel="canonical" href="https://yourbusiness.com/" />
Your content is assembled in the browser. Anything that does not execute JavaScript, and that includes most AI crawlers, receives an empty root div and nothing else. The page is not thin, and it is not low quality. It is blank.
<div id="root"></div>
There is no llms.txt, and robots.txt says nothing about GPTBot, ClaudeBot, PerplexityBot, or OAI-SearchBot. You are not missing from AI answers because you are small. You are missing because nothing on your site ever invited those crawlers in, and what they found when they came was empty.
robots.txt: no rules for GPTBot, ClaudeBot, PerplexityBot
The AI half
When someone asks an assistant to recommend a business like yours, the model is not opening your site and admiring it. It is working from text it collected earlier, from your pages and from the places that write about you. If your pages returned nothing, you are not in the running, however good the design was.
title: Vite + React description: (none) h1: (none) body text: (none) schema: (none) llms.txt: 404
This is the whole record. There is no version of this where an assistant recommends you, because there is nothing here to recommend.
Writing more content will not solve this. What solves it is getting the content you already wrote into the response before any JavaScript runs, and letting the crawlers that read it through the door.
How I get vibe coded sites citedProof, not promises
Every number below is a screenshot from Google Search Console or the Bing AI Performance report, not a rounded claim in a case study I wrote about myself. The first one is a site designed, built, and written with the same AI workflow you used, then pointed at search properly.
481
clicks from launch · 3 months
A new tech site to 481 clicks and a climbing curve in 90 days
208.4K
AI citations · 3 months
208.4K AI citations in three months
23.3K
organic clicks · 6 months
23.3K clicks from 3.55M impressions in six months
All numbers are taken directly from Google Search Console, Google Business Profile, GA4, Bing Webmaster Tools, and the Microsoft AI Performance dashboard for active clients and owned projects. Names withheld at client request.
Three ways to work with me
Every engagement is the same person doing the work, from the first x-ray to the last dashboard. Pick the one that matches where your site is today.
What I write about
I would rather you read the fix and apply it yourself than pay me for something you could have done in an afternoon. The people who want it handled know where to find me.
Canonicals, render modes, routing, sitemaps. The specific failures that come out of Lovable, Cursor, v0, and Bolt, and the exact edit that fixes each one.
How ChatGPT, Perplexity, and AI Overviews decide which businesses to name, what they read to decide it, and which of those sources you can actually influence.
Why a corrected build can sit ignored for months, and the exact steps that make Google come back and look again.
What I find running these checks across real sites, written up with the data attached. Not opinions about the industry, measurements from inside it.
Who is writing this
I build with the same tools you do, across every site I own, which is how I know exactly where they leave you exposed. I work with a small number of clients at a time, because the alternative is handing your account to someone junior and calling it scale.
Questions
A business whose website was built by describing it to an AI tool such as Lovable, Cursor, v0, Bolt, or Replit, usually by the owner, usually in days. It is a new category of business owner: you shipped a site that looks better than most agency work, without ever hiring a developer, and now you need it to be found. That is the only kind of client this site is for.
In almost every case it is one of three things. Every page carries a canonical tag pointing at the homepage, so Google treats the whole site as one URL. Or the page is empty until JavaScript runs, so crawlers that do not run it receive nothing. Or the site was never submitted and has no sitemap worth reading. None of these are design problems, and none of them need a rebuild.
Usually, yes. The failures live in the document head, the routing, the render mode, and the sitemap, not in the design you spent your weekends on. Those are editable. A rebuild is the last resort, and I will tell you honestly if yours is the rare case that needs one.
No. React, Astro, and Next sites rank perfectly well when they are server rendered and their tags are correct. This site is built in Lovable and Claude Code, it is indexed, and its schema is in the raw HTML. The builder is not the problem. The default output is the problem, and defaults can be changed.
Because fixing the build and fixing the index are two different clocks. Once Google has filed your pages as copies of the homepage, it records that decision and stops coming back. Something has to force a re-fetch, page by page. That step is part of every engagement I run, because without it the fix is invisible.
A normal agency starts with keywords and content. Your site cannot use either yet, because search engines have not read it. I start where your site actually is: I x-ray what the crawler receives, fix the reasons it is blank, force the recrawl, and only then move on to ranking and AI citations. I also build sites with the same tools you used, so I know exactly where they leave you exposed.
Work with Ashikur Rahman
Send me the URL. I will x-ray what Google and the AI engines actually receive from it, tell you exactly why it is invisible, and show you what I would do about it. Free, and a real person reads it.
