Originally published on We0 on November 16, 2026. This English edition was adapted with AI assistance.
An AI-built website can be crawled and indexed. What matters is the content returned by the server, what becomes available after JavaScript runs, and whether Google can access and index the page. Server-rendered or statically generated content reduces reliance on a separate rendering step.
Inspect the Raw HTML, Then Compare the Rendered Page
Start by checking the HTTP status and page source. Compare the body text, headings, and links with what appears after rendering. The following commands are quick clues, not a rendering diagnosis by themselves. Set the target URL first:
SITE_URL="https://example.com/product"
curl -sL "$SITE_URL" -o raw.html -w "HTTP %{http_code}\n"
grep -o "<h1[^>]*>[^<]*" raw.html | head -n 5 # Whether there is a real h1
grep -o 'href="[^"]*"' raw.html | head -n 10 # Whether there are followable links
grep -c "<script" raw.html # A rough script count; not proof of CSR or SSR
Next, use URL Inspection in Search Console. After entering the target URL, compare two places: the crawled HTML (the version Google fetched last time), and the rendered result and screenshot obtained after testing the live URL. If the former has no body text while the latter does, the page is highly rendering-dependent; if neither has it, then it truly cannot be read.
Then complete the basics. Submit sitemap.xml and declare it in robots.txt, confirm that JS and CSS are not blocked, and set a self-referencing canonical for indexable pages.
User-agent: *
Allow: /
# For sitemaps, write an absolute address pointing to /sitemap.xml under your own domain
Three Diagnostic Patterns for a Landing Page
After running the three steps above on a marketing landing page generated by an AI website builder, there are commonly three results:
- The raw HTML contains
h1and product descriptions, with a single-digit number of scripts: the page exposes useful content without waiting for JavaScript. This helps discoverability but does not guarantee indexing. - The raw HTML contains only a container node, the
hrefs found bygrepare almost all#, and there are more than a dozen scripts: a pattern worth checking for client-side rendering; script count alone does not establish the architecture. Whether it can be read depends on the rendering process and whether APIs are allowed. - The raw HTML has structure but no body text, while the rendered screenshot shows complete content: the body text comes from client-side requests, and once the API is blocked by robots.txt, supply is immediately cut off.
Three results correspond to three actions: keep the status quo, change the rendering method, or fix robots.txt. Distinguishing which one applies before acting saves more cost than directly switching templates.
When Should You Switch to Server-Side Rendering or Prerendering?
If two or more of the following conditions are met, stop betting visibility on client-side rendering:
- The page mainly carries lead-generation tasks: homepage, product pages, solution pages, pricing pages, case study pages, blog.
- Content is relatively stable and does not depend on login state or real-time personalization.
- The target market includes overseas or multilingual markets, requiring independent URLs and independently indexable pages.
- The team has no dedicated front-end engineer available to troubleshoot rendering issues at any time and wants a one-time configuration that works long term.
- The content needs to be cited in AI search and Q&A scenarios, requiring content to be stably extracted into paragraphs.
Conversely, dashboards after login and highly interactive tool interfaces are perfectly reasonable to build with CSR—such pages should not have been expected to bring traffic from search engines in the first place, and it is more cost-effective to focus effort on the in-product experience.
From 'Can Be Crawled' to 'Can Be Cited'
Indexing is only a ticket for entry. To reach the point of being cited by search engines and generative engines, it depends on extractable information in the page:
- For each core question, give a definition-style direct answer first, then expand the argument; each paragraph should cover only one thing.
- Use tables to carry comparison relationships (rendering methods, solution differences, specifications), which are easier to extract than long paragraphs.
- Attach source links and dates to key numbers, versions, and prices so citations have evidence.
- Keep the spelling of brand, product, and region names consistent to avoid multiple spellings of the same entity.
- Add structured data for FAQs, products, and organizational information to help understand the page's role.
These things are unrelated to the rendering method, but they only take effect on the premise that the body text can be read. First solve 'can be read,' then solve 'worth citing'; the order cannot be reversed.
FAQ
Are websites generated by AI website builders definitely unable to be indexed by Google?
No. Indexing depends on whether there is readable content after rendering, whether crawling is allowed, and whether there are indexable URLs—not on who generated the website.
If the page source contains no body text, will it be judged as cheating?
No. This is not an integrity issue but a technical implementation issue. The criterion is whether Google can obtain the body text after rendering.
Will switching to server-side rendering make me rank at the top?
This should not be understood that way. SSR and prerendering improve content readability and indexing stability; ranking is still determined by content quality, page experience, and competition.
Is dynamic rendering still worth investing in?
It is suitable as a transitional measure during migration, but Google positions dynamic rendering as a workaround, and in the long term it is still recommended to move rendering forward to the server or build stage.
Related Tools
- Google Search Console: A core tool for URL inspection, sitemap submission, and troubleshooting crawl and index status.
- curl: Confirms the raw HTML returned by the server, bypassing browser rendering and cache.
- Browser 'View Page Source': Quickly determines whether body text exists in the first HTML.
- Structured data testing tool: Verifies whether markup such as FAQ, product, and organization can be parsed.
- we0's SEO and GEO optimization panel: Manages page structure and content within the same workspace and checks visibility after launch (we0.ai).
Related Links
- Rendering on the web (web.dev): Definitions, performance trade-offs, and SEO considerations for CSR, SSR, static rendering, and prerendering.
- Understand JavaScript SEO basics (Google Search Central): The basic process and limitations of how Googlebot handles JavaScript pages.
- Dynamic rendering as a workaround (Google Search Central): The positioning and applicable boundaries of dynamic rendering.
- we0.ai: The capability list of the AI website building and lead generation workspace, including the AI website generator, CMS backend, domain deployment, and SEO/GEO optimization.
Summary
Whether websites generated by AI website builders can be normally crawled and indexed by Google does not depend on 'AI-generated' but on the rendering method. Google supports JavaScript rendering, but rendering is an independent stage after crawling and is subject to limits around queuing, resource fetching, and execution boundaries; if the page source contains no body text, visibility is built on an unstable premise. For showcase-oriented sites focused on official websites, landing pages, product pages, and blogs, prioritize server-side rendering or static generation, write the body text into the first HTML, and then complete it with sitemaps, canonical tags, and structured content, so pages move from 'can be read' to 'worth citing.' The verification method is straightforward: inspect the raw HTML, use curl to confirm status codes and titles, and compare the crawled HTML with the rendered result in Search Console.

Top comments (0)