Is AI-generated code bad for SEO?

Not inherently. The risk is that AI coding tools optimize for a working, good-looking result and often skip semantic HTML, metadata, canonical URLs, structured data and crawlable rendering unless you explicitly ask for them.

Does JavaScript hurt SEO?

Not by default, but it adds a step. Google crawls, renders and then indexes a page, and rendering can be delayed, so important content, links and metadata should ideally be present in the initial HTML rather than only appearing after the page fully executes in the browser.

What is GEO and AEO, and do they really affect rankings?

GEO (Generative Engine Optimization) and AEO (Answer Engine Optimization) describe making content easier for generative AI systems to understand and surface. Google explicitly treats these as extensions of normal SEO fundamentals, not as separate ranking hacks.

How do I make my website visible in ChatGPT Search?

Allow OpenAI's OAI-SearchBot in your robots.txt, since sites that opt it out won't appear in ChatGPT's search answers. Allowing it is a precondition for inclusion, not a guarantee that your pages will actually be cited.

Does structured data help AI understand a website?

It helps make the entities, relationships and content type on a page explicit rather than implied, which benefits both traditional rich results and AI systems trying to summarize a page — though it doesn't guarantee any particular placement.

Can you optimize an existing AI-built website instead of rebuilding it?

Usually yes. Most issues — missing metadata, poor heading structure, no sitemap, no structured data, slow rendering — can be fixed incrementally on top of an existing vibe-coded site without a full rebuild.

← Guides

Can a Vibe-Coded Website Rank on Google and AI Search?

3 October 2026

Can a Vibe-Coded Website Rank on Google and Show Up in AI Search?

I can vibe code a website very quickly. The harder question is whether Google, ChatGPT and other systems can actually understand what I built.

A landing page that used to take days or weeks can now be produced in an afternoon with AI. The prompt works, the layout looks good, the code runs. That used to be the hard part. Now it's the easy part — and it's where most AI-built websites stop. Getting that same page crawled, understood, indexed and surfaced is a separate job, with its own rules, and it doesn't happen automatically just because the code is clean.

I've written before about why vibe coding is only the beginning and why vibe coding is only half the job. This article is the practical follow-up: what actually has to be true for a vibe-coded site to be found.

Can AI-generated websites rank in Google?

Yes, in the sense that Google doesn't care how a page was produced. Google's documentation on AI features in Search explains that AI Overviews and AI Mode work by surfacing relevant links, and "the AI features AI Overviews and AI Mode surface relevant links to help people find the information they're looking for quickly and reliably, as well as to help them explore content they may not have discovered before. These features offer unique opportunities for more types of sites to appear." Nothing there is conditioned on how the HTML was written or which tool generated it.

What matters is whether the resulting site meets the same technical, quality and spam requirements every site is held to. A vibe-coded site isn't disadvantaged for being AI-built — but it isn't given a pass either.

Why a beautiful vibe-coded website can still be invisible

I see the same handful of problems over and over in AI-generated sites: content that only appears after JavaScript runs, flat or missing heading structure, duplicated page titles, no canonical URLs, no sitemap, weak internal linking, and no structured data at all. None of these break the page visually. All of them can quietly keep it out of the index.

Google can process JavaScript, but it does so in stages, and that matters for anything built with a modern frontend framework. Google's own explanation of how Googlebot processes JavaScript walks through three phases: "Google processes JavaScript web apps in three main phases: Crawling, Rendering, Indexing. Googlebot takes a URL from the crawl queue, crawls it, then passes it into the processing stage." Rendering isn't instant, either — "Googlebot queues all pages with a 200 HTTP status code for rendering, unless a robots meta tag or header tells Google not to index the page. The page may stay on this queue for a few seconds, but it can take longer than that." That's a real reason server-side rendering or static generation still matters for anything you want indexed quickly, and Google says as much: "server-side or pre-rendering is still a great idea because it makes your website faster for users and crawlers, and not all bots can run JavaScript."

SEO should exist while you build, not after

The framing I keep coming back to: SEO isn't something you bolt on once the site is finished. It's part of the architecture, the same way the database schema or the component structure is.

That means thinking about semantic headings, meaningful HTML, crawlable links, unique titles and descriptions, canonical URLs, an XML sitemap, internal linking and structured data from the first commit — not as a final pass before launch. If the page hierarchy is wrong at the prototype stage, no amount of metadata patching fixes it later.

SEO, GEO and AEO are not magic tricks

SEO is helping search engines discover, understand and index content. GEO is improving the chance that content gets understood and surfaced in generative search experiences. AEO is structuring information so a system can identify a clear answer to a question. None of these are hacks, and Google is explicit about that. Its own resource on optimizing for generative AI in Search includes "Guidance on the importance of providing valuable, unique, non-commodity content... Mythbusting common 'AEO/GEO' misconceptions... Initial guidance related to AI agents, a quickly emerging and evolving space". In other words: the fundamentals are the fundamentals, with or without the new acronyms.

How I build a landing page differently

Before I write code, I think about what the business actually offers, who the page is for, what questions those people are typing, and what information a search engine or an AI system should be able to extract from the page without asking me.

While building, I'm thinking about semantic HTML, a clear page hierarchy, unique metadata per page, canonical URLs, internal links, JSON-LD using Schema.org vocabulary, accessibility, and whether the important content is present in the initial HTML or only after a script runs.

After building, I test: Lighthouse scores, Core Web Vitals, the actual rendered HTML (not just what I see in the browser), the sitemap, robots.txt, and whether AI crawlers can reach the pages at all. This is the same habit I describe in a case where a scroll animation stuttered on phones — measuring the actual cost instead of guessing, rather than assuming a page is fine because it looks fine.

KodeKid as the test case

I use KodeKid itself as a running experiment rather than a brochure. Part of that has meant shipping deliberately rather than defaulting to a full stack on day one — I've documented going live with no backend at all for a stretch, before adding a database, auth or a CMS, specifically so the site could get online and get indexed without unnecessary complexity getting in the way. The point isn't that any particular stack is correct — it's that the technical choices behind a site should be made on purpose, not left as a side effect of whichever AI tool generated the first draft.

What Google needs from a vibe-coded site

A practical checklist I use on every project:

Crawlability — can Googlebot reach the important pages at all?
Indexability — is the page eligible for the index, or blocked by a stray noindex?
Content — is there something useful and original here, not boilerplate?
Architecture — does the structure communicate what the site is about?
Links — can Google find pages through normal crawlable <a href> links?
Metadata — unique titles and descriptions on every page that matters?
Structured data — does JSON-LD clearly describe the entities on the page, the way Google's guidance on structured data recommends?
Rendering — is the content actually present in the rendered HTML?
Performance and accessibility — does it work on a mid-range phone, and can assistive tech parse it?

What AI search needs from a website

AI search systems need the same raw material: clear entity names, direct answers, logical headings, and content that's actually crawlable. For ChatGPT specifically, OpenAI operates a dedicated crawler for this purpose. Its documentation on OpenAI's crawlers states "OAI-SearchBot is for search. OAI-SearchBot is used to surface websites in search results in ChatGPT's search features. Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers". Allowing that bot in robots.txt is a precondition for inclusion, not a guarantee of placement — OpenAI never frames it as the latter, and neither should anyone writing about it.

Landing pages are especially interesting

Most AI-generated landing pages are optimized for one question: does it look good? A search-ready landing page answers a different one: does it make sense — to a visitor, to a crawler, to an AI system summarizing the page in one sentence? That means one clear search intent, one clear offer, an obvious hierarchy, a strong title, supporting internal links, and a fast, workable mobile experience. That gap — between an AI-generated landing page and a search-ready one — is most of what this work actually is.

Can you put a vibe-coded website at the top of Google?

No one can guarantee a #1 ranking — not me, not any tool, not any agency. What can be controlled is the set of conditions that make a site competitive: technical quality, content quality, relevance, crawlability, structured data, internal linking, performance, and ongoing iteration based on what's actually happening in Search Console. I can build and optimize those conditions. I can't control Google's algorithm, and I won't pretend otherwise.

From vibe coding to search-ready product

Vibe coding changed how quickly I can build a website. It didn't change what makes a website discoverable. AI gives me speed. Design gives me direction. Code gives me control. SEO gives the site discoverability. Structured data gives machines context. Accessibility gives the interface a clearer structure. And AI search is making machine-readable content more important, not less.

This is also why I build the design, the frontend, the backend and the search layer myself rather than handing pieces off — an approach I go into in why I build it all myself.

Have an AI-built website that looks good but isn't getting found? I can help turn it into a proper search-ready website or landing page — have a look at how I approach this in more depth in vibe coding is only half the job.

Author: Lars Sohl, founder of Klokly.app and KodeKid

Frequently asked questions

Can a vibe-coded website rank on Google?: Yes — Google doesn't check how a site was built, only whether it meets the same crawling, indexing, content and quality standards as any other site. A vibe-coded site with weak technical SEO will struggle just as a manually coded one would with the same issues.
Is AI-generated code bad for SEO?: Not inherently. The risk is that AI coding tools optimize for a working, good-looking result and often skip semantic HTML, metadata, canonical URLs, structured data and crawlable rendering unless you explicitly ask for them.
Does JavaScript hurt SEO?: Not by default, but it adds a step. Google crawls, renders and then indexes a page, and rendering can be delayed, so important content, links and metadata should ideally be present in the initial HTML rather than only appearing after the page fully executes in the browser.
What is GEO and AEO, and do they really affect rankings?: GEO (Generative Engine Optimization) and AEO (Answer Engine Optimization) describe making content easier for generative AI systems to understand and surface. Google explicitly treats these as extensions of normal SEO fundamentals, not as separate ranking hacks.
How do I make my website visible in ChatGPT Search?: Allow OpenAI's OAI-SearchBot in your robots.txt, since sites that opt it out won't appear in ChatGPT's search answers. Allowing it is a precondition for inclusion, not a guarantee that your pages will actually be cited.
Does structured data help AI understand a website?: It helps make the entities, relationships and content type on a page explicit rather than implied, which benefits both traditional rich results and AI systems trying to summarize a page — though it doesn't guarantee any particular placement.
Can you optimize an existing AI-built website instead of rebuilding it?: Usually yes. Most issues — missing metadata, poor heading structure, no sitemap, no structured data, slow rendering — can be fixed incrementally on top of an existing vibe-coded site without a full rebuild.

ENDA