Updated on 18 November 2026, by João Leite, Founder of murmur.
A site built in Lovable appears in ChatGPT when the search bot receives the finished page text in the code the server delivers. Lovable writes, in its documentation, that an app published there returns readable HTML to search engines and AI bots. In older apps, built in React and Vite, Lovable delivers a pre-rendered version only to the bots it verifies, and warns that content that loads late may be missing. Framer and Wix write that they build the page on the server, and Webflow, that the site comes with semantic HTML.
What each platform’s documentation says is the starting point, and the test that decides applies to all of them: open the page source and look for the text. If the main sentence, the company name and the price are there, the bot receives the page ready. If they are not, the bot receives less than the person sees on screen.
- A site built in Lovable delivers readable HTML, according to Lovable. Apps created since 13 May 2026 build the page on the server; older ones, in React and Vite, send a ready copy to the bots that Lovable verifies.
- Framer and Wix write that they build the page on the server. Framer pre-renders each page in HTML, and Wix uses server-side rendering.
- Webflow writes, on its sales page, that the site comes with semantic HTML and with robots.txt and sitemap out of the box.
- No platform handles the text that comes in later, by script, by widget or inside an image.
- The test is the same on all of them: the page source, with the main sentence, the company name and the price written there.
A site built in Lovable, Framer or Wix: AI reads it when the bot receives the finished text
A site built in Lovable, Framer or Wix is read by AI when the bot receives the finished text, without depending on the browser to build the page. AI bots (crawlers) are the programs AI companies use to read sites, such as OpenAI’s OAI-SearchBot, which fetches pages for ChatGPT. The bot requests the page and reads what the server sends; what only appears later, when the browser runs the code, may not reach it.
That is why the right question is not which platform the company uses, but how each page reaches the bot. Lovable, Framer and Wix write about this in their own documentation, and each gives an answer; Webflow talks about the site’s HTML on its sales page. The reason for having the text ready in the HTML, and the list of bots from each AI company, are in the guide about llms.txt and the text the bot reads.
A site built in Lovable: the new app, with the text built on the server, and the old one, with the copy for verified bots
A site built in Lovable reaches the bot in two ways, depending on the age of the app. In Lovable’s SEO and AI search documentation, accessed on 9 November 2026, the company, which sells a website and app builder with AI, writes that published apps return readable HTML to search engines and AI bots. Those created since 13 May 2026 build the page on the server.
Older apps, built in React and Vite, work differently. Lovable writes that they receive a pre-rendered version when a verified bot requests the page, and the list of those bots belongs to Lovable. On the page about switching to TanStack Start, Lovable describes the old app as one that sends the browser an almost empty page. And the same SEO documentation warns, about the pre-rendered version: “Content that loads late may be missing.”
For anyone with a site built in Lovable, the first step is to know whether the app is new or old, and the second is to open the source of the pages that matter. The page that carries the text in the source arrives ready; the one that carries only the empty structure depends on the copy Lovable sends to verified bots.
A site built in Framer: the page pre-rendered on the server
A site built in Framer reaches the bot as ready HTML, according to Framer, which sells a website builder. In the help center, in an article about making the site readable for AI agents, Framer writes that, although the sites are built in React, each page is pre-rendered in HTML on its servers before being delivered. Accessed on 9 November 2026. Text that an outside script adds later, on this or any platform, calls for the source test.
A site built in Wix: server-side rendering and the warning about the page source
A site built in Wix also arrives at the bot assembled on the server, according to Wix, which sells a website builder. In the help center in Portuguese, Wix writes that its sites use server-side rendering to convert the page code into the version the visitor sees. The same page, accessed on 9 November 2026, teaches how to view the page source the way search engines see it, and carries the warning that applies to any platform: “Qualquer coisa que esteja faltando no código-fonte da página é conteúdo que os rastreadores não podem ler.” (“Anything missing from the page source is content that crawlers cannot read.”)
A site built in Webflow: what Webflow’s sales page says
A site built in Webflow comes with semantic HTML, according to Webflow’s sales page, which sells a website builder. On the sales page about AEO, accessed on 9 November 2026, Webflow writes that every site of theirs comes with semantic HTML, schema markup, SEO metadata, llms.txt, sitemap and robots.txt out of the box. That is what Webflow writes about its own product, and the source of each page remains the test.
Platform by platform: how the page reaches the bot
Platform by platform, each company describes how the page reaches the bot, and all pages were accessed on 9 November 2026.
| platform | how the page reaches the bot, according to the platform | the warning or limit the platform writes | page accessed on |
|---|---|---|---|
| Lovable, app created since 13 May 2026 | built on the server, in readable HTML | robots.txt comes in every new project, and not always the sitemap | 9 November 2026 |
| Lovable, old app in React and Vite | pre-rendered copy for the bots that Lovable verifies | content that loads late may be missing | 9 November 2026 |
| Framer | each page pre-rendered in HTML on the server | the sites are built in React, and the build happens on the server | 9 November 2026 |
| Wix | server-side rendering | what is missing from the source, crawlers do not read | 9 November 2026 |
Lovable, Framer and Wix write that they deliver the page built, each in its own way. Lovable separates the new app from the old one, and Framer and Wix talk about building on the server. In none of them does the documentation replace the source test, which is what shows the page the bot actually receives.
What no platform handles: the text that comes in later, through a script or widget
No platform handles the text that comes into the page later, through a script, a widget or an outside app. The pricing table that comes from an external service, the reviews that load in a frame and the chat that shows the answers only after a click can stay out of the page source, and Wix writes that what is missing there crawlers cannot read. Lovable gives the same warning about content that loads late.
Text written inside an image also stays out. Price, deadline and company name go in text, and not only in the banner. The list of what to check on the site, from text readable without JavaScript to the links between pages, is in the technical checklist to appear in ChatGPT and the other AI models.
How to check a site built in Lovable or on any platform: the page source
To check a site built in Lovable or on any platform, open the source of each page that matters and look for the text. The test takes a few minutes and does not call for a paid tool.
- Open the published page, at the URL the buyer will use.
- Open the page source, with Ctrl+U in Chrome.
- Look for the main sentence of the page with the browser search, Ctrl+F.
- Look for the company name and the price, when the page has a price.
- Check whether the links to the other pages appear in the code as URLs.
- Repeat the test on the home page, on a service or product page and on the frequently asked questions page.
If the sentence is in the source code, the bot receives the page with it. If it is not, the page reaches the bot with less than the person sees, and the person who manages the site needs to change how that part is built. In the old Lovable app, the source code the browser receives is the nearly empty page, and the text depends on the copy that goes to verified bots. The robots.txt and the sitemap are part of the same check, and the llms.txt on each platform has a guide of its own, about llms.txt on WordPress and Shopify.
At Mindo, murmur did not touch the main site: the guide was published on a subdomain that delivers the text ready-made
murmur publishes the content on a subdomain of the brand, which delivers the text ready to the bot, and does not touch the main site. These are “guias num subdomínio da marca do cliente, que a murmur monta e mantém, sem mexer no site principal; o conteúdo continua do cliente depois do contrato” (guides on a subdomain of the client’s brand, which murmur builds and maintains, without touching the main site; the content remains the client’s after the contract). The case study that shows this is Mindo, a B2B motion design studio in São Paulo.
On 29 May 2026, in murmur’s diagnostic, “o site era um aplicativo que monta a página no navegador, e o robô das IAs via uma página vazia” (the site was an application that builds the page in the browser, and the AI bot saw an empty page). murmur published Mindo’s guide at guia.mindo.com.br, fully live on 22 June 2026. On the same questions, ChatGPT recommended Mindo in 23% of the questions in its market on 13 and 14 July 2026, against 12% on 25 and 26 June.
The guide on the subdomain does not change what the main site delivers to the bot. The site that arrives empty still needs fixing, and the source code test shows where the problem is. Whether content on a subdomain appears in AI models like the main site’s is covered in the guide about content on a subdomain.
Frequently asked questions
Do I need to change platforms to appear in ChatGPT?
Not necessarily: what decides is whether the source code of each page carries the text. Lovable, Framer and Wix state that they deliver the page ready to bots, and the source code test shows whether that holds for your site. In the old Lovable app, in React and Vite, Lovable itself has a page about switching to TanStack Start. For the startup building its first site, the GEO steps are in the guide about how to do GEO for startups.
Does the lovable.app address appear in search?
Lovable states that only the published and public app can appear in search results. In the same documentation, it recommends connecting a custom domain to the app to build search presence.
Does Lovable create the robots.txt and the sitemap on its own?
Lovable creates the robots.txt in every new project, but not always the sitemap, according to its documentation. The same page states that the starting metadata, such as title and description, is generic. So it is worth checking the sitemap and replacing the title and description of each page with the buyer’s words.
Is llms.txt enough for the site that builds the page in the browser?
It is not enough: llms.txt points to the pages, and does not change what each page delivers to the bot. llms.txt is a text file, at the site’s main URL, that shows AI models its most important pages. If the page arrives empty, the file leads the bot to an empty page. What is known about the file is in the guide about llms.txt.
Where to start
Start by opening the source code of your home page and looking for its main sentence. Then see what AI models already answer about your market: murmur’s free diagnostic shows, without sign-up and in under two minutes, what AI models answer when someone searches for what your company sells, and the full report goes by email. murmur’s sector studies show which brands AI models recommended in each market murmur measured.
Read also
- llms.txt: What It Is, What It Is For, and Whether It Works to Get Cited by AI
- Technical checklist to appear in ChatGPT and other AI models: what to check on your site
- How to do GEO for startups: the steps to appear in ChatGPT against the leader
- Subdomain Content Appears in AI in 2026: What Changes Compared to the Main Site
References
- Lovable, SEO and AI search documentation, undated page. App published with HTML readable by search engines and AI crawlers; older app built with React and Vite with a pre-rendered version for verified bots (“Content that loads late may be missing.”); only the published and public app appears in search; robots.txt in every new project, sitemap not always. Lovable sells an AI website and app builder. https://docs.lovable.dev/features/seo-aeo (accessed on 9 November 2026).
- Lovable, documentation on switching to TanStack Start, undated page. The older app built with React and Vite sends an almost empty page to the browser. https://docs.lovable.dev/features/upgrade-to-tanstack-start (accessed on 9 November 2026).
- Framer, help center, article on making the site readable for AI agents, undated page. Every page pre-rendered in HTML on the server. Framer sells a website builder. https://www.framer.com/help/articles/make-site-readable-by-ai-agents/ (accessed on 9 November 2026).
- Wix, help center in Portuguese, page on viewing the source code the way search engines see it, undated page. Server-side rendering; “Qualquer coisa que esteja faltando no código-fonte da página é conteúdo que os rastreadores não podem ler.” (“Anything missing from the page’s source code is content that crawlers cannot read.”) Wix sells a website builder. https://support.wix.com/pt/article/visualizar-o-codigo-fonte-da-pagina-da-maneira-que-os-motores-de-busca-fazem (accessed on 9 November 2026).
- Webflow, sales page about AEO, undated page. Semantic HTML, schema, metadata, llms.txt, sitemap and robots.txt out of the box, in Webflow’s words. https://webflow.com/solutions/aeo (accessed on 9 November 2026).