What the homepages of 9 website builders ship: HTML from 123 KB to 2.7 MB, and hreflang from 0 to 28 tags
When a tool reads a site it starts with the HTML of a page. On 9 November 2026 we requested the homepage of nine builders (curl, following redirects, one request each) and counted the size of the HTML, the hreflang attributes, the JSON-LD blocks and the canonical tag. These are the builders' own marketing sites, not customer sites.
What we found
| Platform | HTML | hreflang | JSON-LD | Note |
|---|---|---|---|---|
| Wix | 2.67 MB | 22 | 5 | canonical without trailing slash |
| Framer | 2.12 MB | 3 | 1 | canonical tagged data-framer-generated-canonical |
| Jimdo | 862 KB | 3 | 2 | |
| Squarespace | 626 KB | 24 | 5 | |
| Webflow | 553 KB | 28 | 1 | |
| WordPress.com | 341 KB | 25 | 1 | |
| Ghost | 123 KB | 0 | 1 | canonical attribute written without quotes |
| GoDaddy | 403 | – | – | 369-byte refusal |
| Weebly | 403 | – | – | Cloudflare challenge page |
Three things to take from it
- The HTML is not the text. Wix's homepage is 2.67 MB of HTML and Ghost's is 123 KB, yet both are a single page of marketing copy; we did not measure how much of each is text. Anything that extracts a post has to find the article inside that, which is why one extractor per platform is needed.
- Every readable homepage declares a canonical. They differ in small ways (a trailing slash, a vendor attribute, unquoted values), so a strict parser that expects one exact form will miss some.
- Language tags vary widely. Webflow, WordPress.com, Squarespace and Wix list 22 to 28 hreflang attributes; Ghost has none. When you move a multilingual site, the alternates are part of what you must rebuild.
Limits
One request per site, from one machine, with a research user agent. The hreflang count is the number of hreflang= attributes in the HTML, which can include duplicates. GoDaddy and Weebly returned a 403, so their pages are unknown. Sizes are uncompressed bytes of the HTML, not what the browser downloads over the wire.
Related
See also which robots.txt files name AI crawlers and what the sitemaps look like. Or scan your own site for free before you pay for anything.