The difference is not just whether crawlers execute JavaScript, but whether they can access the main content afterward. In the source article’s comparison, the raw HTML of both a plain Vue single-page application (SPA) and Flutter Web may lack the main content. The Vue SPA produces parseable DOM text after JavaScript executes, while the article’s Flutter Web implementation uses CanvasKit to draw text on a canvas. Vue can use Nuxt SSR to deliver the main content upfront; the Flutter Web implementation described cannot solve the issue simply by waiting for JavaScript rendering. This comparison is limited to the implementations described in the article and does not imply that all versions and architectures behave the same way.
When raw HTML lacks content, first distinguish SSR from CSR
In the Flutter Web implementation tested by the source author, the body of the raw HTML consisted mainly of loading scripts. The Vue comparison project used a plain Vite SPA, whose raw HTML likewise contained only an empty app container. Comparing these with Nuxt SSR’s complete HTML therefore highlights the difference between server-side rendering (SSR) and client-side rendering (CSR). It does not establish that Vue is inherently easier to index than Flutter.
- CSR: The main content appears only after JavaScript executes, so retrieving the initial HTML does not mean retrieving the content.
- Nuxt SSR: HTML containing the main content is generated on the server, allowing crawlers to read the text without waiting for client-side rendering.
According to the article, some search crawlers execute JavaScript after retrieving the HTML, but rendering may still take time. SSR improves how content is delivered; it does not guarantee indexing or rankings.
After rendering, DOM text and canvas content still differ
The Vue SPA in the article produces parseable elements and text after JavaScript executes, while Flutter Web displays its output on a canvas. The author therefore argues that relying on SEO experience with typical Vue CSR implementations would underestimate the risks of this type of Flutter Web implementation: visible text is not necessarily available as parseable DOM content.
Flutter’s semantics tree is another potential source of confusion. It can provide nodes with aria-label attributes for assistive technologies, but the author notes that it is not designed to represent article structure or let crawlers follow links to content. Even forcing it on at startup does not establish that the SEO issues have been resolved. Accessibility support and indexing of the main content still need to be checked separately.
The source article also summarizes Flutter’s official documentation as positioning Flutter Web for applications rather than document- or content-oriented websites. If a product primarily relies on articles or product descriptions to attract search traffic, how that content will be delivered should be determined when choosing the technology.
Per-page metadata matters, but cannot replace the main content
The Nuxt SSR example retrieves article data based on the route, then writes metadata such as the title, summary, and OG image information into the HTML so that preview crawlers that do not execute JavaScript can read it. Updating only shared, site-wide static settings may leave different articles displaying the same information when shared.
In the author’s view, Flutter Web requires more custom work to set route-specific titles and integrate SEO tools. However, not every oversight can be blamed on the framework. Both of the author’s comparison projects retained their default titles, and the Flutter project also retained its default description. These settings should be corrected first. Then check the main content in the raw HTML, each page’s metadata, and the content structure after JavaScript executes separately.
SEO challenges in Flutter Web and Vue: perspectives from across the web
How can you check what content crawlers can actually access?
Start by using curl to inspect the raw HTML returned by the URL and check whether the main content, title, and other information are present. Then use Search Console’s URL Inspection tool to examine the result rendered by Googlebot. Checking these two stages separately lets you distinguish whether content is present initially, appears only after JavaScript executes, or remains inaccessible after rendering. The results reflect that page’s implementation at the time of testing and do not guarantee indexing or rankings.
Why should you check the sharing preview for every internal page?
Updating only the shared, site-wide static title, description, and OG information may cause different internal pages to display identical previews when shared. The source uses Nuxt SSR to demonstrate writing each page’s title and image information into the raw HTML based on its route, allowing preview crawlers that do not execute JavaScript to read it. Test several different internal pages rather than checking only the homepage. A correct preview does not necessarily mean the main content will be indexed.
If the default title and description were left unchanged, is that an SEO problem caused by the framework?
Not necessarily. Both the source author’s Flutter and Vue comparison projects retained their default titles, and the Flutter project’s description was also left unchanged. The author acknowledged this as a personal oversight, not something attributable to the frameworks. Before launch, check and correct the titles and descriptions actually being served, then assess route-level information and delivery of the main content separately to avoid mistaking configuration omissions for inherent framework limitations.
References
- Day 26|Flutter Web 的 SEO 困境:爬蟲看到的是什麼ithelp.ithome.com.tw — Discusses: How can you check what content crawlers can actually access?; Why should you check the sharing preview for every internal page?



