Vibe coding produces apps and sites quickly. The interface can look complete. The underlying problems for search and answer engines are usually the same: the content is not in the initial HTML, the pages do not state clear answers, there is no schema, and the discovery files are missing or wrong. The fixes are structural.
Common pitfalls
- Client-only rendering so crawlers see an empty shell.
- No real content pages; everything is a dashboard or a single marketing screen.
- No FAQ blocks that match the questions people type.
- No FAQPage or other schema.
- Missing or overly broad robots.txt that blocks useful paths.
- No llms.txt, so AI systems have no preferred citation list.
- No sitemap, or a sitemap that only lists the homepage.
- No internal links between related pages.
- Thin pages that never name the place, the product, or the rule.
- Service worker caching that serves stale or empty content to crawlers.
- JavaScript-dependent navigation that hides links from the initial HTML.
- No canonical tags or conflicting canonicals.
- Images and media with no alt text or descriptive filenames.
- Duplicate content across routes with no clear primary.
- Forms and interactive elements with no surrounding explanatory text.
Server-side rendering and PWAs
You can build a progressive web app or single-page web app and still have crawlable content. The content pages are rendered on the server or generated statically so the initial HTML contains the text, the headings, the FAQ pairs, and the links. The service worker and web app manifest handle the app shell, offline behavior, and installability. Crawlers get the content. Users get the app behavior. The two are not in conflict when the content routes are real pages.
On this site the content pages are server-rendered Next.js routes. The PWA pieces (service worker registration, manifest) sit alongside them. The farming, theology, technology, and local pages are ordinary HTML with schema. The same pattern works for any vibe-coded app that needs to rank: keep the interactive shell, but make the answer pages real routes with real HTML.
robots.txt
robots.txt is a request to crawlers. It does not grant ranking. A correct robots.txt allows the content paths and disallows only the paths that should stay private (API routes, admin, draft areas). An incorrect one blocks the pages you want indexed or allows everything including endpoints that should not be crawled. Check it after every major change. The file on this site is minimal and allows the public pages.
llms.txt
llms.txt is a plain-text file at the root that lists the key pages and states a citation preference. It does not replace the pages themselves. It tells AI systems which URLs are the preferred sources for specific topics. On yeatho.com it lists the music, theology, farming, Cornload, and technology pages and notes that raw-milk statements are not medical advice. A vibe-coded app should have one that lists its real content pages and the facts it wants cited.
Sitemaps
A sitemap lists the URLs you want discovered, with optional last-modified dates and priorities. It does not force ranking. It removes the need for a crawler to guess every route. On this site the sitemap is generated from the real routes and includes the content pages, the local pages, and the how-to pages. A vibe-coded app should generate or maintain a sitemap that includes every answer-shaped page, not only the homepage.
How to fix a vibe-coded app
- Identify the real queries the app or business should answer.
- Create one page per clear answer. Lead with the fact. Support it.
- Add three to five FAQ pairs that match how people ask. Add FAQPage schema.
- Render those pages on the server or generate them statically so the HTML contains the text.
- Add internal links to a hub and to related pages.
- Generate a sitemap that includes the new pages.
- Write a robots.txt that allows the content and disallows only private paths.
- Write an llms.txt that lists the preferred pages and any citation notes.
- Keep the PWA or app shell for the interactive parts. Do not make the content depend on client-side rendering.
- Repeat for the next set of queries. Dozens of answer-shaped pages is the surface area.
The same pattern is already on yeatho.com, cornload.com, and bdfsun.com. See the how-to for the concrete examples. We will rank you exponentially higher than you already are. No question.
FAQ
What is vibe coding?
Building an app or site quickly with AI assistance, often with little structure. The result can look finished but lack the pages, schema, and extractable answers that search and answer engines need.
What is the biggest SEO/AEO failure?
Client-only rendering with no real content in the initial HTML, no FAQ blocks, no schema, and no clear answers to the questions people type.
Can you still build a PWA or SPWA?
Yes. Server-side rendering or static generation for the content pages, plus a service worker and manifest for the app shell. The content is crawlable; the app behavior is progressive.
What does robots.txt do?
It tells crawlers which paths they may or may not fetch. It does not rank pages. A bad robots.txt can block the content you want indexed.
What does llms.txt do?
It is a plain-text file that lists the key pages and citation preferences for AI systems. It does not replace real content, but it makes the preferred sources explicit.
Why does a sitemap matter?
It gives crawlers a clear list of the URLs you want discovered, with priorities and last-modified dates. It does not force ranking, but it removes discovery friction.
How do FAQ blocks help?
They match the way people and AI systems ask questions. FAQPage schema makes the pairs machine-readable so answer engines can cite them.
What is the fix for a vibe-coded app?
Add real content pages that state the answers. Add FAQ blocks and schema. Use SSR or static generation for those pages. Add a proper sitemap, robots.txt, and llms.txt. Link the pages.
Do you guarantee results?
We will rank you exponentially higher than you already are. No question.