Schema is Structured Data, which means it is Structured so Google bots know the Structure and can crawl it faster to get an idea of page content faster…it’s pretty essential to Great SEO
2 Likes
I think a simple text can be done. If you go to our Bubble website, pick any page and right click > view source. Do you see the page text in that source? You will likely see the index header text etc, but in all my tests there is no page content text. This is the issue.
You need to look on the Elements tab - not the Source tab. As others have said, the lion’s share of a Bubble page is generated client-side. Google does execute JS much like a browser, so the crawler will see the final rendered page.
Also, the content in the Page HTML Header field is evaluated / generated server-side - hence the constraints on what can be referenced.
That is interesting; I asked various AI bots if they can crawl my website and they said no - they asked me to check the Source tab and then basically said “if you can’t see the text in your source tab, neither can we”.
I then emailed Bubble support and their Bot replied and confirmed this. I am unsure what to believe now!
Not all crawlers do, but Google does, as well as some others. Unfortunately, I don’t have an exhaustive list of which ones do. Maybe consult an LLM. 
I did some tests on this recently.
Google’s standard search engine does pick up bubble site content ok, even dynamic content such as blog posts - as long as the sitemap and unique SLUGS are done correctly.
But all LLMs - even Gemini by Google, fail at reading bubble sites correctly. They can do a very good job at pretending they do, but when pushed on details they admit they can’t.
As we move more and more towards a world where the Ai will replace traditional search, I think bubble really needs to figure this out - fast. It is its potential biggest downfall. 
Are bubble aware of this??
I have been looking into ways to serve static html pages. It’s not actually possible with prerender if you’re on the shared cluster, but I believe it IS possible with Cloudflare workers. Tests still ongoing. The thing is we shouldn’t “really” be having to resort to such extreme workarounds. Wish someone at bubble would acknowledge and address this.
1 Like
Yes @asked111, you’re identifying the same issues I’ve been seeing. I’ve also been testing and exploring potential workarounds. In fact, I’m seriously considering developing a solution. The reality is that most Bubble users won’t know how to implement these changes, or even realize that their sites aren’t being properly indexed.
While Prerender isn’t designed solely for Bubble, I believe we need something similar that is. As you mentioned, LLMs are quickly becoming the default search engines, so any effective solution should be able to remove hidden AI elements, generate AI-first schema, quickly create slugs, sitemaps and more.
Yes totally agree. I believe @boston85719 is also working on a possibly solution for this.
@fede.bubble is there anyway you can find out if there is an ‘official word’ from bubble on this important aspect? If bubble.io sites can not be read by a.i crawlers that is going to be a major turn of for potential users so definitely requires to be taken seriously.
Ok great, the more SEO solutions available for users the better. I have emailed Bubble and will update here when I get a reply.
Believe the humans with real experience on the subject, not bots that are trained to regurgitate information.
Google bots crawl and index bubble apps setup properly and have been for years. AI bots can crawl bubble apps too, just not all of them. It’s in my view, more of an AI issue, and not really Bubble. You can check forum threads where it’s discussed at length.
Bubble already does everything needed for traditional SEO, I’m not sure if there is anything for them to do for AI to crawl a rendered page.
2 Likes
I have tested myself for days!
- Google sometimes renders JavaScript to see the DOM
- AI crawlers (ChatGPT, Perplexity) usually don’t render JavaScript
- It’s unreliable
This is a clearer overview of how you can test yourself:
- View Page Source
-
Shows the raw HTML sent by the server.
-
Most AI models and basic crawlers can only see this.
-
Bubble apps often have little to no content here, because content loads via JavaScript.
-
- Inspect → Elements (Rendered DOM)
-
Shows the HTML after JavaScript has executed.
-
Google and Bing can see and index this content, because they render JavaScript.
-
LLMs (like ChatGPT without browsing) cannot see this, because they do not execute JS.
-
1 Like
I would just like to add that chatGPT can definitely, 100% crawl my Bubble page.
I know because I just asked it “tell me about X” and it picked up text that wasn’t in any structured data/markup, just regular old text in a “text box.” No robots.txt customization or anything special.
So it’s absolutely false that “AI crawlers cannot see Bubble pages.”
2 Likes
Interesting. Hope that is correct 
My guess is that it may totally depend on how fast the page is loaded…if it can load fast enough, the ai may hang around long enough to read it. So with bubble, different times of the day may wield different success rates.
I will do more extensive tests today using a mix of static & dynamic text. Will share the test page link & results here.
Yes that is true but only partially, let me explain;
What you experienced:
-
ChatGPT Browsed you page via Google’s index
-
Google rendered the JavaScript and indexed it
-
ChatGPT then read Google’s cached version
The problems:
-
This only works IF Google already indexed it (takes weeks/months). Some pages may not be indexed.
-
Perplexity, SearchGPT crawl directly - they DON’T see the content
-
It’s unreliable - Google doesn’t always render JS properly
1 Like
Correct. Google’s crawler has limited time to wait for JS, so slow or resource-heavy Bubble pages may not fully render before the crawler times out.
1 Like
To be fair, it was static text, not dynamic.
I’m doing comprehensive testing on AI search visibility for Bubble apps and wanted to share what I’m finding.
While some Bubble pages do get picked up by ChatGPT (via Google’s index), I’m discovering this is inconsistent and depends on several factors: page load speed, how quickly Google crawls/indexes the site, and whether AI tools are using Google’s index vs. crawling directly.
I’m testing all scenarios - different URL structures, page types, crawlers (ChatGPT, Perplexity, SearchGPT), and various Bubble configurations.
The goal is to identify what works reliably vs. what works sometimes.I’m spending a few weeks on this to find a 100% reliable solution that ensures Bubble content is immediately visible to all AI search engines, not just hoping Google indexes it eventually. Will share findings as I go.
I’d love to hear anyone’s thoughts and experiences; what’s working for you, what isn’t, and any patterns you’ve noticed. We’re all figuring this out!
1 Like
Hi,
Google manage to index JS pages but it would be better with HTML.
It is possible to index bubble pages on Google in less than 48h by submitting your sitemap directly in the Google Search Console.
Ab Advani from the community has build a tool https://getpagerabbit.com/ (formerly known as CoAlias) to cache bubble pages in HTML format in order to make them “as readable” as a “native HTML pages” by Google crawlers.
“Bubble apps face SEO challenges due to dynamic content and client-side rendering. PageRabbit caches dynamic pages as static versions for crawlers, boosting visibility simply.”
You can check the live he did during the last "DevDay” in march : https://www.youtube.com/watch?v=-y9EHf0lk-o
1 Like
Thanks for sharing. You’re absolutely right that Google can index JS pages, but HTML is definitely optimal for SEO performance.
I am building a solution that addresses this exact problem. The tool I am working on automatically converts your JavaScript-heavy pages into crawler-accessible HTML format, ensuring search engines can properly index your content without the overhead of rendering JavaScript.
Key benefits:
-
Content becomes immediately visible to crawlers (no JS execution needed)
-
Improved indexing speed and reliability
-
Complete meta tags and robots directives optimization
-
Auto-adds AI schema markup for enhanced discoverability
-
Add additional formatted text content in AI-preferred structures (question-answer format) for better visibility in AI search engines
Seems similar to PageRabbit’s approach, but with meta tag completion, crawler accessibility checks, AI optimization, and ongoing optimization actions. I am about 30% through the build, I will keep you posted here. I am tested over and over as I go, and if it does exactly what I hope I will be a happy man!
2 Likes